This is an archive of the discontinued LLVM Phabricator instance.

Differential D157273

[GlobalISel] Add dead flags to implicit defs in ISel
ClosedPublic

Authored by Pierre-vh on Aug 7 2023, 4:00 AM.

Details

Reviewers
aemerson
arsenm
foad
RKSimon
tschuett
Commits
rGc3cfbbc4160c: [GlobalISel] Add dead flags to implicit defs in ISel
Summary

Checks for implicit defs that are unused within a pattern and mark them as dead.

This is done directly at the TableGen level forr efficiency.
The instructions are directly created with the "dead" operand and no further analysis is needed later.

Diff Detail

Event Timeline

Pierre-vh created this revision.Aug 7 2023, 4:00 AM
Herald added a project: Restricted Project. · View Herald TranscriptAug 7 2023, 4:00 AM
Herald added subscribers: StephenFan, kerbowa, pengfei and 4 others. · View Herald Transcript
Pierre-vh requested review of this revision.Aug 7 2023, 4:00 AM
Herald added a project: Restricted Project. · View Herald TranscriptAug 7 2023, 4:00 AM
Herald added subscribers: llvm-commits, wdng. · View Herald Transcript
Pierre-vh added a comment.Aug 7 2023, 4:02 AM

A couple of things to note:

  • This is a bit naive, it only works on the Record* and doesn't test for subregs or RC, should it?
  • One of the tradeoffs of doing this at the TableGen level is that we don't know about reserved registers. Do they matter in this case? How could we handle them?
foad added a comment.Aug 7 2023, 4:04 AM

Are there any cases where this improves codegen?

Pierre-vh added a comment.Aug 7 2023, 4:08 AM
In D157273#4565239, @foad wrote:

Are there any cases where this improves codegen?

The only codegen change I spotted was in mul-scalar.ll (X86)
The intent is just to improve correctness. @arsenm noticed that not setting dead flags was an issue in https://github.com/llvm/llvm-project/issues/63986

Harbormaster completed remote builds in B250719: Diff 547706.Aug 7 2023, 5:22 AM
arsenm added a comment.Aug 7 2023, 9:37 AM
In D157273#4565254, @Pierre-vh wrote:
In D157273#4565239, @foad wrote:

Are there any cases where this improves codegen?

The only codegen change I spotted was in mul-scalar.ll (X86)
The intent is just to improve correctness. @arsenm noticed that not setting dead flags was an issue in https://github.com/llvm/llvm-project/issues/63986

It's not really a correctness fix, it's just avoiding regressing > 30 globalisel tests when my fix for that is applied. Most of the cases we're compensating for missing constant folding in GlobalISel

llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h
335

Can just make the flag value?

335

actually you can compact the two into one field, register is 32-bits and the flags are 16 I think

390–392

Unrelated formatting change

llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutorImpl.h
951–952

I'd expect the table entry to just be any flags an unconditionally or'd

arsenm added a comment.Aug 7 2023, 2:15 PM
In D157273#4565233, @Pierre-vh wrote:

A couple of things to note:

  • This is a bit naive, it only works on the Record* and doesn't test for subregs or RC, should it?

Probably can ignore this

  • One of the tradeoffs of doing this at the TableGen level is that we don't know about reserved registers. Do they matter in this case? How could we handle them?

It doesn't matter, dead flags shouldn't be set on reserved registers but also shouldn't matter

Pierre-vh updated this revision to Diff 548132.Aug 8 2023, 2:48 AM
Pierre-vh marked 4 inline comments as done.

Comments

Pierre-vh added a comment.Aug 8 2023, 2:49 AM
In D157273#4567283, @arsenm wrote:
In D157273#4565233, @Pierre-vh wrote:

A couple of things to note:

  • This is a bit naive, it only works on the Record* and doesn't test for subregs or RC, should it?

Probably can ignore this

  • One of the tradeoffs of doing this at the TableGen level is that we don't know about reserved registers. Do they matter in this case? How could we handle them?

It doesn't matter, dead flags shouldn't be set on reserved registers but also shouldn't matter

I guess if we want to be 100% correct we could check for the presence of RegState::Dead in the flags, and check if we're dealing with a reserved register.
If it doesn't matter though let's leave it at that, it can always be added later.

llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h
335

I think compacting it is more trouble than it's worth, and it'd make this opcode an outlier - e.g. AddTempRegister doesn't compact, AddTempSubRegister doesn't, etc.

We should just look into making this a variable-length encoded array at some point, but it's quite a bit of work. Probably not worth it until we get actual performance numbers.

I wonder if we ever make use of the upper 32 bits though, maybe the table should always be 32 bits, with an option to variable-length-encode bigger values.

arsenm accepted this revision.Aug 8 2023, 4:17 AM
arsenm added inline comments.
llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h
335

could change all these easy cases at once

llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutorImpl.h
950–951

This could have just been in the table

This revision is now accepted and ready to land.Aug 8 2023, 4:17 AM
Harbormaster completed remote builds in B251044: Diff 548132.Aug 8 2023, 5:01 AM
Closed by commit rGc3cfbbc4160c: [GlobalISel] Add dead flags to implicit defs in ISel (authored by Pierre-vh). · Explain WhyAug 9 2023, 5:21 AM
This revision was automatically updated to reflect the committed changes.
Pierre-vh added a commit: rGc3cfbbc4160c: [GlobalISel] Add dead flags to implicit defs in ISel.

Revision Contents

PathSize
llvm/
include/
llvm/
CodeGen/
GlobalISel/
4 lines
4 lines
test/
CodeGen/
AArch64/
GlobalISel/
8 lines
4 lines
2 lines
AMDGPU/
GlobalISel/
17 lines
28 lines
52 lines
6 lines
18 lines
10 lines
49 lines
34 lines
52 lines
24 lines
15 lines
16 lines
11 lines
8 lines
8 lines
38 lines
52 lines
2 lines
2 lines
9 lines
2 lines
2 lines
24 lines
26 lines
8 lines
8 lines
26 lines
13 lines
4 lines
22 lines
19 lines
4 lines
12 lines
8 lines
8 lines
26 lines
13 lines
4 lines
22 lines
19 lines
4 lines
12 lines
16 lines
4 lines
30 lines
24 lines
22 lines
16 lines
4 lines
30 lines
24 lines
22 lines
202 lines
8 lines
8 lines
26 lines
14 lines
4 lines
4 lines
13 lines
4 lines
4 lines
8 lines
8 lines
26 lines
14 lines
4 lines
4 lines
13 lines
4 lines
4 lines
14 lines
4 lines
14 lines
4 lines
Mips/
GlobalISel/
instruction-select/
55 lines
352 lines
X86/
GlobalISel/
2 lines
4 lines
48 lines
6 lines
4 lines
36 lines
48 lines
20 lines
48 lines
utils/
TableGen/
28 lines
59 lines
3 lines

Diff 548574

llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h

Show First 20 Lines • Show All 326 Lines • ▼ Show 20 Linesenum {
/// - OldInsnID - Instruction ID to copy from /// - OldInsnID - Instruction ID to copy from
/// - OpIdx - The operand to copy /// - OpIdx - The operand to copy
/// - SubRegIdx - The subregister to copy /// - SubRegIdx - The subregister to copy
GIR_CopySubReg, GIR_CopySubReg,
/// Add an implicit register def to the specified instruction /// Add an implicit register def to the specified instruction
/// - InsnID - Instruction ID to modify /// - InsnID - Instruction ID to modify
/// - RegNum - The register to add /// - RegNum - The register to add
/// - Flags - Register Flags
arsenmUnsubmitted
Done
ReplyInline Actions

Can just make the flag value?

arsenm: Can just make the flag value?
arsenmUnsubmitted
Done
ReplyInline Actions

actually you can compact the two into one field, register is 32-bits and the flags are 16 I think

arsenm: actually you can compact the two into one field, register is 32-bits and the flags are 16 I…
Pierre-vhAuthorUnsubmitted
Done
ReplyInline Actions

I think compacting it is more trouble than it's worth, and it'd make this opcode an outlier - e.g. AddTempRegister doesn't compact, AddTempSubRegister doesn't, etc.

We should just look into making this a variable-length encoded array at some point, but it's quite a bit of work. Probably not worth it until we get actual performance numbers.

I wonder if we ever make use of the upper 32 bits though, maybe the table should always be 32 bits, with an option to variable-length-encode bigger values.

Pierre-vh: I think compacting it is more trouble than it's worth, and it'd make this opcode an outlier - e.
arsenmUnsubmitted
Not Done
ReplyInline Actions

could change all these easy cases at once

arsenm: could change all these easy cases at once
GIR_AddImplicitDef, GIR_AddImplicitDef,
/// Add an implicit register use to the specified instruction /// Add an implicit register use to the specified instruction
/// - InsnID - Instruction ID to modify /// - InsnID - Instruction ID to modify
/// - RegNum - The register to add /// - RegNum - The register to add
GIR_AddImplicitUse, GIR_AddImplicitUse,
/// Add an register to the specified instruction /// Add an register to the specified instruction
/// - InsnID - Instruction ID to modify /// - InsnID - Instruction ID to modify
/// - RegNum - The register to add /// - RegNum - The register to add
Show All 38 Linesenum {
/// - InsnID - Instruction ID to modify /// - InsnID - Instruction ID to modify
/// - OldInsnID - Instruction ID to get the matched operand from /// - OldInsnID - Instruction ID to get the matched operand from
/// - RendererFnID - Custom renderer function to call /// - RendererFnID - Custom renderer function to call
GIR_CustomRenderer, GIR_CustomRenderer,
/// Calls a C++ function to perform an action when a match is complete. /// Calls a C++ function to perform an action when a match is complete.
/// The MatcherState is passed to the function to allow it to modify /// The MatcherState is passed to the function to allow it to modify
/// instructions. /// instructions.
/// This is less constrained than a custom renderer and can update instructions /// This is less constrained than a custom renderer and can update
/// instructions
/// in the state. /// in the state.
arsenmUnsubmitted
Done
ReplyInline Actions

Unrelated formatting change

arsenm: Unrelated formatting change
/// - FnID - The function to call. /// - FnID - The function to call.
/// TODO: Remove this at some point when combiners aren't reliant on it. It's /// TODO: Remove this at some point when combiners aren't reliant on it. It's
/// a bit of a hack. /// a bit of a hack.
GIR_CustomAction, GIR_CustomAction,
/// Render operands to the specified instruction using a custom function, /// Render operands to the specified instruction using a custom function,
/// reading from a specific operand. /// reading from a specific operand.
/// - InsnID - Instruction ID to modify /// - InsnID - Instruction ID to modify
▲ Show 20 Lines • Show All 207 Lines • Show Last 20 Lines

llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutorImpl.h

Show First 20 Lines • Show All 939 Lines • ▼ Show 20 Lines case GIR_CopySubReg: {
<< NewInsnID << "], MIs[" << OldInsnID << "], " << NewInsnID << "], MIs[" << OldInsnID << "], "
<< OpIdx << ", " << SubRegIdx << ")\n"); << OpIdx << ", " << SubRegIdx << ")\n");
break; break;
} }
case GIR_AddImplicitDef: { case GIR_AddImplicitDef: {
int64_t InsnID = MatchTable[CurrentIdx++]; int64_t InsnID = MatchTable[CurrentIdx++];
int64_t RegNum = MatchTable[CurrentIdx++]; int64_t RegNum = MatchTable[CurrentIdx++];
auto Flags = (uint64_t)MatchTable[CurrentIdx++];
assert(OutMIs[InsnID] && "Attempted to add to undefined instruction"); assert(OutMIs[InsnID] && "Attempted to add to undefined instruction");
OutMIs[InsnID].addDef(RegNum, RegState::Implicit); Flags |= RegState::Implicit;
OutMIs[InsnID].addDef(RegNum, Flags);
arsenmUnsubmitted
Not Done
ReplyInline Actions

This could have just been in the table

arsenm: This could have just been in the table
DEBUG_WITH_TYPE(TgtExecutor::getName(), DEBUG_WITH_TYPE(TgtExecutor::getName(),
arsenmUnsubmitted
Done
ReplyInline Actions

I'd expect the table entry to just be any flags an unconditionally or'd

arsenm: I'd expect the table entry to just be any flags an unconditionally or'd
dbgs() << CurrentIdx << ": GIR_AddImplicitDef(OutMIs[" dbgs() << CurrentIdx << ": GIR_AddImplicitDef(OutMIs["
<< InsnID << "], " << RegNum << ")\n"); << InsnID << "], " << RegNum << ")\n");
break; break;
} }
case GIR_AddImplicitUse: { case GIR_AddImplicitUse: {
int64_t InsnID = MatchTable[CurrentIdx++]; int64_t InsnID = MatchTable[CurrentIdx++];
int64_t RegNum = MatchTable[CurrentIdx++]; int64_t RegNum = MatchTable[CurrentIdx++];
▲ Show 20 Lines • Show All 269 Lines • Show Last 20 Lines

llvm/test/CodeGen/AArch64/GlobalISel/opt-fold-compare.mir

Show First 20 Lines • Show All 99 Lines • ▼ Show 20 Linesbody: |
bb.0: bb.0:
liveins: $w0, $w1 liveins: $w0, $w1
; CHECK-LABEL: name: no_cmn_s32_rhs ; CHECK-LABEL: name: no_cmn_s32_rhs
; CHECK: liveins: $w0, $w1 ; CHECK: liveins: $w0, $w1
; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0 ; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0
; CHECK: [[COPY1:%[0-9]+]]:gpr32 = COPY $w1 ; CHECK: [[COPY1:%[0-9]+]]:gpr32 = COPY $w1
; CHECK: [[COPY2:%[0-9]+]]:gpr32 = COPY $wzr ; CHECK: [[COPY2:%[0-9]+]]:gpr32 = COPY $wzr
; CHECK: [[SUBSWrr:%[0-9]+]]:gpr32 = SUBSWrr [[COPY2]], [[COPY1]], implicit-def $nzcv ; CHECK: [[SUBSWrr:%[0-9]+]]:gpr32 = SUBSWrr [[COPY2]], [[COPY1]], implicit-def dead $nzcv
; CHECK: [[SUBSWrr1:%[0-9]+]]:gpr32 = SUBSWrr [[COPY]], [[SUBSWrr]], implicit-def $nzcv ; CHECK: [[SUBSWrr1:%[0-9]+]]:gpr32 = SUBSWrr [[COPY]], [[SUBSWrr]], implicit-def $nzcv
; CHECK: [[CSINCWr:%[0-9]+]]:gpr32 = CSINCWr [[COPY2]], $wzr, 10, implicit $nzcv ; CHECK: [[CSINCWr:%[0-9]+]]:gpr32 = CSINCWr [[COPY2]], $wzr, 10, implicit $nzcv
; CHECK: $w0 = COPY [[CSINCWr]] ; CHECK: $w0 = COPY [[CSINCWr]]
; CHECK: RET_ReallyLR implicit $w0 ; CHECK: RET_ReallyLR implicit $w0
%0:gpr(s32) = COPY $w0 %0:gpr(s32) = COPY $w0
%1:gpr(s32) = COPY $w1 %1:gpr(s32) = COPY $w1
%2:gpr(s32) = G_CONSTANT i32 0 %2:gpr(s32) = G_CONSTANT i32 0
%6:gpr(s32) = G_CONSTANT i32 1 %6:gpr(s32) = G_CONSTANT i32 1
Show All 14 Linesbody: |
bb.0: bb.0:
liveins: $w0, $w1 liveins: $w0, $w1
; CHECK-LABEL: name: no_cmn_s32_lhs ; CHECK-LABEL: name: no_cmn_s32_lhs
; CHECK: liveins: $w0, $w1 ; CHECK: liveins: $w0, $w1
; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0 ; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0
; CHECK: [[COPY1:%[0-9]+]]:gpr32 = COPY $w1 ; CHECK: [[COPY1:%[0-9]+]]:gpr32 = COPY $w1
; CHECK: [[COPY2:%[0-9]+]]:gpr32 = COPY $wzr ; CHECK: [[COPY2:%[0-9]+]]:gpr32 = COPY $wzr
; CHECK: [[SUBSWrr:%[0-9]+]]:gpr32 = SUBSWrr [[COPY2]], [[COPY]], implicit-def $nzcv ; CHECK: [[SUBSWrr:%[0-9]+]]:gpr32 = SUBSWrr [[COPY2]], [[COPY]], implicit-def dead $nzcv
; CHECK: [[SUBSWrr1:%[0-9]+]]:gpr32 = SUBSWrr [[SUBSWrr]], [[COPY1]], implicit-def $nzcv ; CHECK: [[SUBSWrr1:%[0-9]+]]:gpr32 = SUBSWrr [[SUBSWrr]], [[COPY1]], implicit-def $nzcv
; CHECK: [[CSINCWr:%[0-9]+]]:gpr32 = CSINCWr [[COPY2]], $wzr, 10, implicit $nzcv ; CHECK: [[CSINCWr:%[0-9]+]]:gpr32 = CSINCWr [[COPY2]], $wzr, 10, implicit $nzcv
; CHECK: $w0 = COPY [[CSINCWr]] ; CHECK: $w0 = COPY [[CSINCWr]]
; CHECK: RET_ReallyLR implicit $w0 ; CHECK: RET_ReallyLR implicit $w0
%0:gpr(s32) = COPY $w0 %0:gpr(s32) = COPY $w0
%1:gpr(s32) = COPY $w1 %1:gpr(s32) = COPY $w1
%2:gpr(s32) = G_CONSTANT i32 0 %2:gpr(s32) = G_CONSTANT i32 0
%6:gpr(s32) = G_CONSTANT i32 1 %6:gpr(s32) = G_CONSTANT i32 1
▲ Show 20 Lines • Show All 74 Lines • ▼ Show 20 Linesbody: |
bb.0: bb.0:
liveins: $x0, $x1 liveins: $x0, $x1
; CHECK-LABEL: name: no_cmn_s64_rhs ; CHECK-LABEL: name: no_cmn_s64_rhs
; CHECK: liveins: $x0, $x1 ; CHECK: liveins: $x0, $x1
; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0 ; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0
; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY $x1 ; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY $x1
; CHECK: [[COPY2:%[0-9]+]]:gpr64 = COPY $xzr ; CHECK: [[COPY2:%[0-9]+]]:gpr64 = COPY $xzr
; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY2]], [[COPY1]], implicit-def $nzcv ; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY2]], [[COPY1]], implicit-def dead $nzcv
; CHECK: [[SUBSXrr1:%[0-9]+]]:gpr64 = SUBSXrr [[COPY]], [[SUBSXrr]], implicit-def $nzcv ; CHECK: [[SUBSXrr1:%[0-9]+]]:gpr64 = SUBSXrr [[COPY]], [[SUBSXrr]], implicit-def $nzcv
; CHECK: [[CSINCXr:%[0-9]+]]:gpr64 = CSINCXr [[COPY2]], $xzr, 10, implicit $nzcv ; CHECK: [[CSINCXr:%[0-9]+]]:gpr64 = CSINCXr [[COPY2]], $xzr, 10, implicit $nzcv
; CHECK: $x0 = COPY [[CSINCXr]] ; CHECK: $x0 = COPY [[CSINCXr]]
; CHECK: RET_ReallyLR implicit $x0 ; CHECK: RET_ReallyLR implicit $x0
%0:gpr(s64) = COPY $x0 %0:gpr(s64) = COPY $x0
%1:gpr(s64) = COPY $x1 %1:gpr(s64) = COPY $x1
%2:gpr(s64) = G_CONSTANT i64 0 %2:gpr(s64) = G_CONSTANT i64 0
%6:gpr(s64) = G_CONSTANT i64 1 %6:gpr(s64) = G_CONSTANT i64 1
Show All 14 Linesbody: |
bb.0: bb.0:
liveins: $x0, $x1 liveins: $x0, $x1
; CHECK-LABEL: name: no_cmn_s64_lhs ; CHECK-LABEL: name: no_cmn_s64_lhs
; CHECK: liveins: $x0, $x1 ; CHECK: liveins: $x0, $x1
; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0 ; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0
; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY $x1 ; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY $x1
; CHECK: [[COPY2:%[0-9]+]]:gpr64 = COPY $xzr ; CHECK: [[COPY2:%[0-9]+]]:gpr64 = COPY $xzr
; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY2]], [[COPY]], implicit-def $nzcv ; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY2]], [[COPY]], implicit-def dead $nzcv
; CHECK: [[SUBSXrr1:%[0-9]+]]:gpr64 = SUBSXrr [[SUBSXrr]], [[COPY1]], implicit-def $nzcv ; CHECK: [[SUBSXrr1:%[0-9]+]]:gpr64 = SUBSXrr [[SUBSXrr]], [[COPY1]], implicit-def $nzcv
; CHECK: [[CSINCXr:%[0-9]+]]:gpr64 = CSINCXr [[COPY2]], $xzr, 10, implicit $nzcv ; CHECK: [[CSINCXr:%[0-9]+]]:gpr64 = CSINCXr [[COPY2]], $xzr, 10, implicit $nzcv
; CHECK: $x0 = COPY [[CSINCXr]] ; CHECK: $x0 = COPY [[CSINCXr]]
; CHECK: RET_ReallyLR implicit $x0 ; CHECK: RET_ReallyLR implicit $x0
%0:gpr(s64) = COPY $x0 %0:gpr(s64) = COPY $x0
%1:gpr(s64) = COPY $x1 %1:gpr(s64) = COPY $x1
%2:gpr(s64) = G_CONSTANT i64 0 %2:gpr(s64) = G_CONSTANT i64 0
%6:gpr(s64) = G_CONSTANT i64 1 %6:gpr(s64) = G_CONSTANT i64 1
▲ Show 20 Lines • Show All 335 Lines • Show Last 20 Lines

llvm/test/CodeGen/AArch64/GlobalISel/select-binop.mir

Show First 20 Lines • Show All 293 Lines • ▼ Show 20 Lines
body: | body: |
bb.0: bb.0:
liveins: $w0, $w1 liveins: $w0, $w1
; CHECK-LABEL: name: sub_s32_gpr ; CHECK-LABEL: name: sub_s32_gpr
; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0 ; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0
; CHECK: [[COPY1:%[0-9]+]]:gpr32 = COPY $w1 ; CHECK: [[COPY1:%[0-9]+]]:gpr32 = COPY $w1
; CHECK: [[SUBSWrr:%[0-9]+]]:gpr32 = SUBSWrr [[COPY]], [[COPY1]], implicit-def $nzcv ; CHECK: [[SUBSWrr:%[0-9]+]]:gpr32 = SUBSWrr [[COPY]], [[COPY1]], implicit-def dead $nzcv
; CHECK: $w0 = COPY [[SUBSWrr]] ; CHECK: $w0 = COPY [[SUBSWrr]]
%0(s32) = COPY $w0 %0(s32) = COPY $w0
%1(s32) = COPY $w1 %1(s32) = COPY $w1
%2(s32) = G_SUB %0, %1 %2(s32) = G_SUB %0, %1
$w0 = COPY %2(s32) $w0 = COPY %2(s32)
... ...
--- ---
Show All 9 Lines
body: | body: |
bb.0: bb.0:
liveins: $x0, $x1 liveins: $x0, $x1
; CHECK-LABEL: name: sub_s64_gpr ; CHECK-LABEL: name: sub_s64_gpr
; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0 ; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0
; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY $x1 ; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY $x1
; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY]], [[COPY1]], implicit-def $nzcv ; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY]], [[COPY1]], implicit-def dead $nzcv
; CHECK: $x0 = COPY [[SUBSXrr]] ; CHECK: $x0 = COPY [[SUBSXrr]]
%0(s64) = COPY $x0 %0(s64) = COPY $x0
%1(s64) = COPY $x1 %1(s64) = COPY $x1
%2(s64) = G_SUB %0, %1 %2(s64) = G_SUB %0, %1
$x0 = COPY %2(s64) $x0 = COPY %2(s64)
... ...
--- ---
▲ Show 20 Lines • Show All 842 Lines • Show Last 20 Lines

llvm/test/CodeGen/AArch64/GlobalISel/select-ptr-add.mir

Show First 20 Lines • Show All 114 Lines • ▼ Show 20 Lines
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $x0, $x1 liveins: $x0, $x1
; CHECK-LABEL: name: ptr_add_negated_reg ; CHECK-LABEL: name: ptr_add_negated_reg
; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0 ; CHECK: [[COPY:%[0-9]+]]:gpr64 = COPY $x0
; CHECK: %src:gpr64 = COPY $x1 ; CHECK: %src:gpr64 = COPY $x1
; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY]], %src, implicit-def $nzcv ; CHECK: [[SUBSXrr:%[0-9]+]]:gpr64 = SUBSXrr [[COPY]], %src, implicit-def dead $nzcv
; CHECK: $x0 = COPY [[SUBSXrr]] ; CHECK: $x0 = COPY [[SUBSXrr]]
%0:gpr(p0) = COPY $x0 %0:gpr(p0) = COPY $x0
%src:gpr(s64) = COPY $x1 %src:gpr(s64) = COPY $x1
%1:gpr(s64) = G_CONSTANT i64 0 %1:gpr(s64) = G_CONSTANT i64 0
%neg:gpr(s64) = G_SUB %1, %src %neg:gpr(s64) = G_SUB %1, %src
%2:gpr(p0) = G_PTR_ADD %0, %neg(s64) %2:gpr(p0) = G_PTR_ADD %0, %neg(s64)
$x0 = COPY %2(p0) $x0 = COPY %2(p0)
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-add.mir

Show All 15 Lines bb.0:
; GFX6-LABEL: name: add_s32 ; GFX6-LABEL: name: add_s32
; GFX6: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4 ; GFX6: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY2]], [[S_ADD_I32_]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY2]], [[S_ADD_I32_]], 0, implicit $exec
; GFX6-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[S_ADD_I32_]], [[V_ADD_CO_U32_e64_]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[S_ADD_I32_]], [[V_ADD_CO_U32_e64_]], 0, implicit $exec
; GFX6-NEXT: [[V_ADD_CO_U32_e64_4:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_5:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_2]], [[COPY2]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_4:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_5:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_2]], [[COPY2]], 0, implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]], implicit [[V_ADD_CO_U32_e64_]], implicit [[V_ADD_CO_U32_e64_2]], implicit [[V_ADD_CO_U32_e64_4]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]], implicit [[V_ADD_CO_U32_e64_]], implicit [[V_ADD_CO_U32_e64_2]], implicit [[V_ADD_CO_U32_e64_4]]
;
; GFX9-LABEL: name: add_s32 ; GFX9-LABEL: name: add_s32
; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4 ; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY2]], [[S_ADD_I32_]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY2]], [[S_ADD_I32_]], 0, implicit $exec
; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[S_ADD_I32_]], [[V_ADD_U32_e64_]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[S_ADD_I32_]], [[V_ADD_U32_e64_]], 0, implicit $exec
; GFX9-NEXT: [[V_ADD_U32_e64_2:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_ADD_U32_e64_1]], [[COPY2]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_2:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_ADD_U32_e64_1]], [[COPY2]], 0, implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]], implicit [[V_ADD_U32_e64_]], implicit [[V_ADD_U32_e64_1]], implicit [[V_ADD_U32_e64_2]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]], implicit [[V_ADD_U32_e64_]], implicit [[V_ADD_U32_e64_1]], implicit [[V_ADD_U32_e64_2]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:vgpr(s32) = COPY $vgpr0 %2:vgpr(s32) = COPY $vgpr0
%3:vgpr(p1) = COPY $vgpr3_vgpr4 %3:vgpr(p1) = COPY $vgpr3_vgpr4
%4:sgpr(s32) = G_CONSTANT i32 1 %4:sgpr(s32) = G_CONSTANT i32 1
%5:sgpr(s32) = G_CONSTANT i32 4096 %5:sgpr(s32) = G_CONSTANT i32 4096
; add ss
%6:sgpr(s32) = G_ADD %0, %1 %6:sgpr(s32) = G_ADD %0, %1
; add vs
%7:vgpr(s32) = G_ADD %2, %6 %7:vgpr(s32) = G_ADD %2, %6
; add sv
%8:vgpr(s32) = G_ADD %6, %7 %8:vgpr(s32) = G_ADD %6, %7
; add vv
%9:vgpr(s32) = G_ADD %8, %2 %9:vgpr(s32) = G_ADD %8, %2
S_ENDPGM 0, implicit %6, implicit %7, implicit %8, implicit %9 S_ENDPGM 0, implicit %6, implicit %7, implicit %8, implicit %9
... ...
--- ---
name: add_neg_inline_const_64_to_sub_s32_s name: add_neg_inline_const_64_to_sub_s32_s
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0 liveins: $sgpr0
; GFX6-LABEL: name: add_neg_inline_const_64_to_sub_s32_s ; GFX6-LABEL: name: add_neg_inline_const_64_to_sub_s32_s
; GFX6: liveins: $sgpr0 ; GFX6: liveins: $sgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], 64, implicit-def $scc ; GFX6-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], 64, implicit-def $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_SUB_I32_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_SUB_I32_]]
;
; GFX9-LABEL: name: add_neg_inline_const_64_to_sub_s32_s ; GFX9-LABEL: name: add_neg_inline_const_64_to_sub_s32_s
; GFX9: liveins: $sgpr0 ; GFX9: liveins: $sgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], 64, implicit-def $scc ; GFX9-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], 64, implicit-def $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_SUB_I32_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_SUB_I32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = G_CONSTANT i32 -64 %1:sgpr(s32) = G_CONSTANT i32 -64
Show All 13 Lines bb.0:
liveins: $vgpr0 liveins: $vgpr0
; GFX6-LABEL: name: add_neg_inline_const_64_to_sub_s32_v ; GFX6-LABEL: name: add_neg_inline_const_64_to_sub_s32_v
; GFX6: liveins: $vgpr0 ; GFX6: liveins: $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_SUB_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[COPY]], 64, 0, implicit $exec ; GFX6-NEXT: [[V_SUB_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[COPY]], 64, 0, implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_SUB_CO_U32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_SUB_CO_U32_e64_]]
;
; GFX9-LABEL: name: add_neg_inline_const_64_to_sub_s32_v ; GFX9-LABEL: name: add_neg_inline_const_64_to_sub_s32_v
; GFX9: liveins: $vgpr0 ; GFX9: liveins: $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_SUB_U32_e64_:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[COPY]], 64, 0, implicit $exec ; GFX9-NEXT: [[V_SUB_U32_e64_:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[COPY]], 64, 0, implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_SUB_U32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_SUB_U32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = G_CONSTANT i32 -64 %1:vgpr(s32) = G_CONSTANT i32 -64
Show All 12 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0 liveins: $sgpr0
; GFX6-LABEL: name: add_neg_inline_const_16_to_sub_s32_s ; GFX6-LABEL: name: add_neg_inline_const_16_to_sub_s32_s
; GFX6: liveins: $sgpr0 ; GFX6: liveins: $sgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[S_MOV_B32_]], implicit-def $scc ; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]]
;
; GFX9-LABEL: name: add_neg_inline_const_16_to_sub_s32_s ; GFX9-LABEL: name: add_neg_inline_const_16_to_sub_s32_s
; GFX9: liveins: $sgpr0 ; GFX9: liveins: $sgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX9-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX9-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[S_MOV_B32_]], implicit-def $scc ; GFX9-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = G_CONSTANT i32 16 %1:sgpr(s32) = G_CONSTANT i32 16
%2:sgpr(s32) = G_ADD %0, %1 %2:sgpr(s32) = G_ADD %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
Show All 9 Lines bb.0:
; GFX6-LABEL: name: add_neg_inline_const_16_to_sub_s32_v ; GFX6-LABEL: name: add_neg_inline_const_16_to_sub_s32_v
; GFX6: liveins: $vgpr0 ; GFX6: liveins: $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 16, implicit $exec ; GFX6-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 16, implicit $exec
; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_]]
;
; GFX9-LABEL: name: add_neg_inline_const_16_to_sub_s32_v ; GFX9-LABEL: name: add_neg_inline_const_16_to_sub_s32_v
; GFX9: liveins: $vgpr0 ; GFX9: liveins: $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 16, implicit $exec ; GFX9-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 16, implicit $exec
; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD_U32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD_U32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = G_CONSTANT i32 16 %1:vgpr(s32) = G_CONSTANT i32 16
%2:vgpr(s32) = G_ADD %0, %1 %2:vgpr(s32) = G_ADD %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-and.mir

Show All 16 Lines bb.0:
; WAVE64-LABEL: name: and_s1_vcc_vcc_vcc ; WAVE64-LABEL: name: and_s1_vcc_vcc_vcc
; WAVE64: liveins: $vgpr0, $vgpr1 ; WAVE64: liveins: $vgpr0, $vgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec ; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]]
; ;
; WAVE32-LABEL: name: and_s1_vcc_vcc_vcc ; WAVE32-LABEL: name: and_s1_vcc_vcc_vcc
; WAVE32: liveins: $vgpr0, $vgpr1 ; WAVE32: liveins: $vgpr0, $vgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec ; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(s32) = G_CONSTANT i32 0 %2:vgpr(s32) = G_CONSTANT i32 0
%3:vcc(s1) = G_ICMP intpred(eq), %0, %2 %3:vcc(s1) = G_ICMP intpred(eq), %0, %2
%4:vcc(s1) = G_ICMP intpred(eq), %1, %2 %4:vcc(s1) = G_ICMP intpred(eq), %1, %2
%5:vcc(s1) = G_AND %3, %4 %5:vcc(s1) = G_AND %3, %4
S_ENDPGM 0, implicit %5 S_ENDPGM 0, implicit %5
▲ Show 20 Lines • Show All 109 Lines • ▼ Show 20 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; WAVE64-LABEL: name: and_s32_sgpr_sgpr_sgpr ; WAVE64-LABEL: name: and_s32_sgpr_sgpr_sgpr
; WAVE64: liveins: $sgpr0, $sgpr1 ; WAVE64: liveins: $sgpr0, $sgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]]
; ;
; WAVE32-LABEL: name: and_s32_sgpr_sgpr_sgpr ; WAVE32-LABEL: name: and_s32_sgpr_sgpr_sgpr
; WAVE32: liveins: $sgpr0, $sgpr1 ; WAVE32: liveins: $sgpr0, $sgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_AND %0, %1 %2:sgpr(s32) = G_AND %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: and_s64_sgpr_sgpr_sgpr name: and_s64_sgpr_sgpr_sgpr
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE64-LABEL: name: and_s64_sgpr_sgpr_sgpr ; WAVE64-LABEL: name: and_s64_sgpr_sgpr_sgpr
; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 ; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]]
; ;
; WAVE32-LABEL: name: and_s64_sgpr_sgpr_sgpr ; WAVE32-LABEL: name: and_s64_sgpr_sgpr_sgpr
; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 ; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
; WAVE32-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s64) = COPY $sgpr2_sgpr3 %1:sgpr(s64) = COPY $sgpr2_sgpr3
%2:sgpr(s64) = G_AND %0, %1 %2:sgpr(s64) = G_AND %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 198 Lines • ▼ Show 20 Lines bb.0:
; WAVE64: liveins: $vgpr0, $vgpr1 ; WAVE64: liveins: $vgpr0, $vgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]]
; ;
; WAVE32-LABEL: name: and_s1_vcc_copy_to_vcc ; WAVE32-LABEL: name: and_s1_vcc_copy_to_vcc
; WAVE32: liveins: $vgpr0, $vgpr1 ; WAVE32: liveins: $vgpr0, $vgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(s1) = G_TRUNC %0 %2:vgpr(s1) = G_TRUNC %0
%3:vgpr(s1) = G_TRUNC %1 %3:vgpr(s1) = G_TRUNC %1
%4:vcc(s1) = COPY %2 %4:vcc(s1) = COPY %2
%5:vcc(s1) = COPY %3 %5:vcc(s1) = COPY %3
%6:vcc(s1) = G_AND %4, %5 %6:vcc(s1) = G_AND %4, %5
Show All 17 Lines bb.0:
; WAVE64: liveins: $vgpr0, $sgpr0 ; WAVE64: liveins: $vgpr0, $sgpr0
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_AND_B64_]] ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_AND_B64_]]
; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]]
; ;
; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave32 ; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
; WAVE32: liveins: $vgpr0, $sgpr0 ; WAVE32: liveins: $vgpr0, $sgpr0
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_AND_B32_1]] ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_AND_B32_1]]
; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]]
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%0:vgpr(s1) = G_TRUNC %1(s32) %0:vgpr(s1) = G_TRUNC %1(s32)
%sgpr0:sgpr(s32) = COPY $sgpr0 %sgpr0:sgpr(s32) = COPY $sgpr0
%2:sgpr(s1) = G_TRUNC %sgpr0 %2:sgpr(s1) = G_TRUNC %sgpr0
%6:sgpr(s32) = G_CONSTANT i32 0 %6:sgpr(s32) = G_CONSTANT i32 0
%7:sgpr(p1) = G_IMPLICIT_DEF %7:sgpr(p1) = G_IMPLICIT_DEF
Show All 19 Lines bb.0:
; WAVE64: liveins: $vgpr0, $sgpr0 ; WAVE64: liveins: $vgpr0, $sgpr0
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B64_]]
; ;
; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave64 ; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
; WAVE32: liveins: $vgpr0, $sgpr0 ; WAVE32: liveins: $vgpr0, $sgpr0
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_AND_B32_1]] ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_AND_B32_1]]
; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]]
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%0:vgpr(s1) = G_TRUNC %1(s32) %0:vgpr(s1) = G_TRUNC %1(s32)
%sgpr0:sgpr(s32) = COPY $sgpr0 %sgpr0:sgpr(s32) = COPY $sgpr0
%2:sgpr(s1) = G_TRUNC %sgpr0 %2:sgpr(s1) = G_TRUNC %sgpr0
%6:sgpr(s32) = G_CONSTANT i32 0 %6:sgpr(s32) = G_CONSTANT i32 0
%7:sgpr(p1) = G_IMPLICIT_DEF %7:sgpr(p1) = G_IMPLICIT_DEF
Show All 15 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; WAVE64-LABEL: name: and_s32_sgpr_sgpr_sgpr_result_reg_class ; WAVE64-LABEL: name: and_s32_sgpr_sgpr_sgpr_result_reg_class
; WAVE64: liveins: $sgpr0, $sgpr1 ; WAVE64: liveins: $sgpr0, $sgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]]
; ;
; WAVE32-LABEL: name: and_s32_sgpr_sgpr_sgpr_result_reg_class ; WAVE32-LABEL: name: and_s32_sgpr_sgpr_sgpr_result_reg_class
; WAVE32: liveins: $sgpr0, $sgpr1 ; WAVE32: liveins: $sgpr0, $sgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_AND_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sreg_32(s32) = G_AND %0, %1 %2:sreg_32(s32) = G_AND %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-ashr.mir

Show All 13 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX6-LABEL: name: ashr_s32_ss ; GFX6-LABEL: name: ashr_s32_ss
; GFX6: liveins: $sgpr0, $sgpr1 ; GFX6: liveins: $sgpr0, $sgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]]
;
; GFX7-LABEL: name: ashr_s32_ss ; GFX7-LABEL: name: ashr_s32_ss
; GFX7: liveins: $sgpr0, $sgpr1 ; GFX7: liveins: $sgpr0, $sgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX7-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX7-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX7-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]]
;
; GFX8-LABEL: name: ashr_s32_ss ; GFX8-LABEL: name: ashr_s32_ss
; GFX8: liveins: $sgpr0, $sgpr1 ; GFX8: liveins: $sgpr0, $sgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]]
;
; GFX9-LABEL: name: ashr_s32_ss ; GFX9-LABEL: name: ashr_s32_ss
; GFX9: liveins: $sgpr0, $sgpr1 ; GFX9: liveins: $sgpr0, $sgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]]
;
; GFX10-LABEL: name: ashr_s32_ss ; GFX10-LABEL: name: ashr_s32_ss
; GFX10: liveins: $sgpr0, $sgpr1 ; GFX10: liveins: $sgpr0, $sgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_ASHR_I32_:%[0-9]+]]:sreg_32 = S_ASHR_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_ASHR %0, %1 %2:sgpr(s32) = G_ASHR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: ashr_s32_sv name: ashr_s32_sv
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; GFX6-LABEL: name: ashr_s32_sv ; GFX6-LABEL: name: ashr_s32_sv
; GFX6: liveins: $sgpr0, $vgpr0 ; GFX6: liveins: $sgpr0, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX7-LABEL: name: ashr_s32_sv ; GFX7-LABEL: name: ashr_s32_sv
; GFX7: liveins: $sgpr0, $vgpr0 ; GFX7: liveins: $sgpr0, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX8-LABEL: name: ashr_s32_sv ; GFX8-LABEL: name: ashr_s32_sv
; GFX8: liveins: $sgpr0, $vgpr0 ; GFX8: liveins: $sgpr0, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX9-LABEL: name: ashr_s32_sv ; GFX9-LABEL: name: ashr_s32_sv
; GFX9: liveins: $sgpr0, $vgpr0 ; GFX9: liveins: $sgpr0, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX10-LABEL: name: ashr_s32_sv ; GFX10-LABEL: name: ashr_s32_sv
; GFX10: liveins: $sgpr0, $vgpr0 ; GFX10: liveins: $sgpr0, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
Show All 12 Lines bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; GFX6-LABEL: name: ashr_s32_vs ; GFX6-LABEL: name: ashr_s32_vs
; GFX6: liveins: $sgpr0, $vgpr0 ; GFX6: liveins: $sgpr0, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX7-LABEL: name: ashr_s32_vs ; GFX7-LABEL: name: ashr_s32_vs
; GFX7: liveins: $sgpr0, $vgpr0 ; GFX7: liveins: $sgpr0, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX8-LABEL: name: ashr_s32_vs ; GFX8-LABEL: name: ashr_s32_vs
; GFX8: liveins: $sgpr0, $vgpr0 ; GFX8: liveins: $sgpr0, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX9-LABEL: name: ashr_s32_vs ; GFX9-LABEL: name: ashr_s32_vs
; GFX9: liveins: $sgpr0, $vgpr0 ; GFX9: liveins: $sgpr0, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX10-LABEL: name: ashr_s32_vs ; GFX10-LABEL: name: ashr_s32_vs
; GFX10: liveins: $sgpr0, $vgpr0 ; GFX10: liveins: $sgpr0, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 12 Lines bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
; GFX6-LABEL: name: ashr_s32_vv ; GFX6-LABEL: name: ashr_s32_vv
; GFX6: liveins: $vgpr0, $vgpr1 ; GFX6: liveins: $vgpr0, $vgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX6-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX7-LABEL: name: ashr_s32_vv ; GFX7-LABEL: name: ashr_s32_vv
; GFX7: liveins: $vgpr0, $vgpr1 ; GFX7: liveins: $vgpr0, $vgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX8-LABEL: name: ashr_s32_vv ; GFX8-LABEL: name: ashr_s32_vv
; GFX8: liveins: $vgpr0, $vgpr1 ; GFX8: liveins: $vgpr0, $vgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX9-LABEL: name: ashr_s32_vv ; GFX9-LABEL: name: ashr_s32_vv
; GFX9: liveins: $vgpr0, $vgpr1 ; GFX9: liveins: $vgpr0, $vgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
;
; GFX10-LABEL: name: ashr_s32_vv ; GFX10-LABEL: name: ashr_s32_vv
; GFX10: liveins: $vgpr0, $vgpr1 ; GFX10: liveins: $vgpr0, $vgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX10-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_ASHRREV_I32_e64_:%[0-9]+]]:vgpr_32 = V_ASHRREV_I32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $sgpr2 liveins: $sgpr0_sgpr1, $sgpr2
; GFX6-LABEL: name: ashr_s64_ss ; GFX6-LABEL: name: ashr_s64_ss
; GFX6: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX6: liveins: $sgpr0_sgpr1, $sgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]]
;
; GFX7-LABEL: name: ashr_s64_ss ; GFX7-LABEL: name: ashr_s64_ss
; GFX7: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX7: liveins: $sgpr0_sgpr1, $sgpr2
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX7-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX7-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]]
;
; GFX8-LABEL: name: ashr_s64_ss ; GFX8-LABEL: name: ashr_s64_ss
; GFX8: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0_sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]]
;
; GFX9-LABEL: name: ashr_s64_ss ; GFX9-LABEL: name: ashr_s64_ss
; GFX9: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0_sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]]
;
; GFX10-LABEL: name: ashr_s64_ss ; GFX10-LABEL: name: ashr_s64_ss
; GFX10: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX10: liveins: $sgpr0_sgpr1, $sgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX10-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_ASHR_I64_:%[0-9]+]]:sreg_64 = S_ASHR_I64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_ASHR_I64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s32) = COPY $sgpr2 %1:sgpr(s32) = COPY $sgpr2
%2:sgpr(s64) = G_ASHR %0, %1 %2:sgpr(s64) = G_ASHR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: ashr_s64_sv name: ashr_s64_sv
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $vgpr0 liveins: $sgpr0_sgpr1, $vgpr0
; GFX6-LABEL: name: ashr_s64_sv ; GFX6-LABEL: name: ashr_s64_sv
; GFX6: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX6: liveins: $sgpr0_sgpr1, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]]
;
; GFX7-LABEL: name: ashr_s64_sv ; GFX7-LABEL: name: ashr_s64_sv
; GFX7: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX7: liveins: $sgpr0_sgpr1, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]]
;
; GFX8-LABEL: name: ashr_s64_sv ; GFX8-LABEL: name: ashr_s64_sv
; GFX8: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0_sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
;
; GFX9-LABEL: name: ashr_s64_sv ; GFX9-LABEL: name: ashr_s64_sv
; GFX9: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0_sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
;
; GFX10-LABEL: name: ashr_s64_sv ; GFX10-LABEL: name: ashr_s64_sv
; GFX10: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX10: liveins: $sgpr0_sgpr1, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
Show All 12 Lines bb.0:
liveins: $sgpr0, $vgpr0_vgpr1 liveins: $sgpr0, $vgpr0_vgpr1
; GFX6-LABEL: name: ashr_s64_vs ; GFX6-LABEL: name: ashr_s64_vs
; GFX6: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX6: liveins: $sgpr0, $vgpr0_vgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]]
;
; GFX7-LABEL: name: ashr_s64_vs ; GFX7-LABEL: name: ashr_s64_vs
; GFX7: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX7: liveins: $sgpr0, $vgpr0_vgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]]
;
; GFX8-LABEL: name: ashr_s64_vs ; GFX8-LABEL: name: ashr_s64_vs
; GFX8: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX8: liveins: $sgpr0, $vgpr0_vgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
;
; GFX9-LABEL: name: ashr_s64_vs ; GFX9-LABEL: name: ashr_s64_vs
; GFX9: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX9: liveins: $sgpr0, $vgpr0_vgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
;
; GFX10-LABEL: name: ashr_s64_vs ; GFX10-LABEL: name: ashr_s64_vs
; GFX10: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX10: liveins: $sgpr0, $vgpr0_vgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
%0:vgpr(s64) = COPY $vgpr0_vgpr1 %0:vgpr(s64) = COPY $vgpr0_vgpr1
Show All 12 Lines bb.0:
liveins: $vgpr0_vgpr1, $vgpr2 liveins: $vgpr0_vgpr1, $vgpr2
; GFX6-LABEL: name: ashr_s64_vv ; GFX6-LABEL: name: ashr_s64_vv
; GFX6: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX6-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]]
;
; GFX7-LABEL: name: ashr_s64_vv ; GFX7-LABEL: name: ashr_s64_vv
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX7-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_ASHR_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHR_I64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_ASHR_I64_e64_]]
;
; GFX8-LABEL: name: ashr_s64_vv ; GFX8-LABEL: name: ashr_s64_vv
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
;
; GFX9-LABEL: name: ashr_s64_vv ; GFX9-LABEL: name: ashr_s64_vv
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
;
; GFX10-LABEL: name: ashr_s64_vv ; GFX10-LABEL: name: ashr_s64_vv
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX10-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_ASHRREV_I64_e64_:%[0-9]+]]:vreg_64 = V_ASHRREV_I64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_ASHRREV_I64_e64_]]
%0:vgpr(s64) = COPY $vgpr0_vgpr1 %0:vgpr(s64) = COPY $vgpr0_vgpr1
%1:vgpr(s32) = COPY $vgpr2 %1:vgpr(s32) = COPY $vgpr2
%2:vgpr(s64) = G_ASHR %0, %1 %2:vgpr(s64) = G_ASHR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-brcond.mir

Show First 20 Lines • Show All 238 Lines • ▼ Show 20 Linesbody: |
; GCN-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GCN-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GCN-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GCN-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GCN-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GCN-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GCN-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[COPY1]], implicit $exec ; GCN-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[COPY1]], implicit $exec
; GCN-NEXT: [[V_CMP_EQ_F32_e64_:%[0-9]+]]:sreg_64_xexec = nofpexcept V_CMP_EQ_F32_e64 0, [[COPY2]], 0, [[COPY3]], 0, implicit $mode, implicit $exec ; GCN-NEXT: [[V_CMP_EQ_F32_e64_:%[0-9]+]]:sreg_64_xexec = nofpexcept V_CMP_EQ_F32_e64 0, [[COPY2]], 0, [[COPY3]], 0, implicit $mode, implicit $exec
; GCN-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_F32_e64_]], implicit-def $scc ; GCN-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_F32_e64_]], implicit-def dead $scc
; GCN-NEXT: $vcc = COPY [[S_AND_B64_]] ; GCN-NEXT: $vcc = COPY [[S_AND_B64_]]
; GCN-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc ; GCN-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: bb.1: ; GCN-NEXT: bb.1:
bb.0: bb.0:
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 22 Linesbody: |
; GCN-NEXT: liveins: $sgpr0, $vgpr0, $vgpr1 ; GCN-NEXT: liveins: $sgpr0, $vgpr0, $vgpr1
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GCN-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GCN-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GCN-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, [[COPY2]], implicit-def dead $scc ; GCN-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, [[COPY2]], implicit-def dead $scc
; GCN-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; GCN-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; GCN-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[COPY1]], implicit $exec ; GCN-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[COPY1]], implicit $exec
; GCN-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_NE_U32_e64_]], implicit-def $scc ; GCN-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_NE_U32_e64_]], implicit-def dead $scc
; GCN-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64 = S_AND_B64 [[S_AND_B64_]], $exec, implicit-def dead $scc ; GCN-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64 = S_AND_B64 [[S_AND_B64_]], $exec, implicit-def dead $scc
; GCN-NEXT: $vcc = COPY [[S_AND_B64_1]] ; GCN-NEXT: $vcc = COPY [[S_AND_B64_1]]
; GCN-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc ; GCN-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: bb.1: ; GCN-NEXT: bb.1:
bb.0: bb.0:
liveins: $sgpr0, $vgpr0, $vgpr1 liveins: $sgpr0, $vgpr0, $vgpr1
Show All 21 Linesbody: |
; GCN: bb.0: ; GCN: bb.0:
; GCN-NEXT: successors: %bb.1(0x80000000) ; GCN-NEXT: successors: %bb.1(0x80000000)
; GCN-NEXT: liveins: $vgpr0, $vgpr1 ; GCN-NEXT: liveins: $vgpr0, $vgpr1
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GCN-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[COPY1]], implicit $exec ; GCN-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[COPY1]], implicit $exec
; GCN-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 -1 ; GCN-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 -1
; GCN-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_EQ_U32_e64_]], [[S_MOV_B64_]], implicit-def $scc ; GCN-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_EQ_U32_e64_]], [[S_MOV_B64_]], implicit-def dead $scc
; GCN-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[S_XOR_B64_]], $exec, implicit-def dead $scc ; GCN-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64 = S_AND_B64 [[S_XOR_B64_]], $exec, implicit-def dead $scc
; GCN-NEXT: $vcc = COPY [[S_AND_B64_]] ; GCN-NEXT: $vcc = COPY [[S_AND_B64_]]
; GCN-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc ; GCN-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: bb.1: ; GCN-NEXT: bb.1:
bb.0: bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
Show All 11 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-build-vector-trunc.v2s16.mir

Show First 20 Lines • Show All 61 Lines • ▼ Show 20 Lines bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX9-LABEL: name: test_build_vector_trunc_s_pack_hl ; GFX9-LABEL: name: test_build_vector_trunc_s_pack_hl
; GFX9: liveins: $sgpr0, $sgpr1 ; GFX9: liveins: $sgpr0, $sgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX9-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX9-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GFX9-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[S_LSHR_B32_]], [[COPY]] ; GFX9-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[S_LSHR_B32_]], [[COPY]]
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]]
; ;
; GFX11-LABEL: name: test_build_vector_trunc_s_pack_hl ; GFX11-LABEL: name: test_build_vector_trunc_s_pack_hl
; GFX11: liveins: $sgpr0, $sgpr1 ; GFX11: liveins: $sgpr0, $sgpr1
; GFX11-NEXT: {{ $}} ; GFX11-NEXT: {{ $}}
; GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
▲ Show 20 Lines • Show All 250 Lines • ▼ Show 20 Lines bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_lh_multi_use ; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_lh_multi_use
; GFX9PLUS: liveins: $sgpr0, $sgpr1 ; GFX9PLUS: liveins: $sgpr0, $sgpr1
; GFX9PLUS-NEXT: {{ $}} ; GFX9PLUS-NEXT: {{ $}}
; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9PLUS-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[COPY]], [[S_LSHR_B32_]] ; GFX9PLUS-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[COPY]], [[S_LSHR_B32_]]
; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]], implicit [[S_LSHR_B32_]] ; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]], implicit [[S_LSHR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_CONSTANT i32 16 %2:sgpr(s32) = G_CONSTANT i32 16
%3:sgpr(s32) = G_LSHR %1, %2 %3:sgpr(s32) = G_LSHR %1, %2
%4:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %0, %3 %4:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %0, %3
S_ENDPGM 0, implicit %4, implicit %3 S_ENDPGM 0, implicit %4, implicit %3
Show All 10 Lines bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_hh_multi_use_lhs ; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_hh_multi_use_lhs
; GFX9PLUS: liveins: $sgpr0, $sgpr1 ; GFX9PLUS: liveins: $sgpr0, $sgpr1
; GFX9PLUS-NEXT: {{ $}} ; GFX9PLUS-NEXT: {{ $}}
; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[S_MOV_B32_]], implicit-def $scc ; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9PLUS-NEXT: [[S_PACK_LH_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LH_B32_B16 [[S_LSHR_B32_]], [[COPY1]] ; GFX9PLUS-NEXT: [[S_PACK_LH_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LH_B32_B16 [[S_LSHR_B32_]], [[COPY1]]
; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LH_B32_B16_]], implicit [[S_LSHR_B32_]] ; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LH_B32_B16_]], implicit [[S_LSHR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_CONSTANT i32 16 %2:sgpr(s32) = G_CONSTANT i32 16
%3:sgpr(s32) = G_LSHR %0, %2 %3:sgpr(s32) = G_LSHR %0, %2
%4:sgpr(s32) = G_LSHR %1, %2 %4:sgpr(s32) = G_LSHR %1, %2
%5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %3, %4 %5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %3, %4
Show All 11 Lines bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX9-LABEL: name: test_build_vector_trunc_s_pack_hh_multi_use_rhs ; GFX9-LABEL: name: test_build_vector_trunc_s_pack_hh_multi_use_rhs
; GFX9: liveins: $sgpr0, $sgpr1 ; GFX9: liveins: $sgpr0, $sgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX9-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX9-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[S_MOV_B32_]], implicit-def $scc ; GFX9-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9-NEXT: [[S_LSHR_B32_1:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GFX9-NEXT: [[S_LSHR_B32_1:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[S_LSHR_B32_]], [[S_LSHR_B32_1]] ; GFX9-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[S_LSHR_B32_]], [[S_LSHR_B32_1]]
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]], implicit [[S_LSHR_B32_1]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]], implicit [[S_LSHR_B32_1]]
; ;
; GFX11-LABEL: name: test_build_vector_trunc_s_pack_hh_multi_use_rhs ; GFX11-LABEL: name: test_build_vector_trunc_s_pack_hh_multi_use_rhs
; GFX11: liveins: $sgpr0, $sgpr1 ; GFX11: liveins: $sgpr0, $sgpr1
; GFX11-NEXT: {{ $}} ; GFX11-NEXT: {{ $}}
; GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX11-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; GFX11-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; GFX11-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GFX11-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX11-NEXT: [[S_PACK_HL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_HL_B32_B16 [[COPY]], [[S_LSHR_B32_]] ; GFX11-NEXT: [[S_PACK_HL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_HL_B32_B16 [[COPY]], [[S_LSHR_B32_]]
; GFX11-NEXT: S_ENDPGM 0, implicit [[S_PACK_HL_B32_B16_]], implicit [[S_LSHR_B32_]] ; GFX11-NEXT: S_ENDPGM 0, implicit [[S_PACK_HL_B32_B16_]], implicit [[S_LSHR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_CONSTANT i32 16 %2:sgpr(s32) = G_CONSTANT i32 16
%3:sgpr(s32) = G_LSHR %0, %2 %3:sgpr(s32) = G_LSHR %0, %2
%4:sgpr(s32) = G_LSHR %1, %2 %4:sgpr(s32) = G_LSHR %1, %2
%5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %3, %4 %5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %3, %4
Show All 11 Lines bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_lh_wrong_shift_amt ; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_lh_wrong_shift_amt
; GFX9PLUS: liveins: $sgpr0, $sgpr1 ; GFX9PLUS: liveins: $sgpr0, $sgpr1
; GFX9PLUS-NEXT: {{ $}} ; GFX9PLUS-NEXT: {{ $}}
; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 15 ; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 15
; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9PLUS-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[COPY]], [[S_LSHR_B32_]] ; GFX9PLUS-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[COPY]], [[S_LSHR_B32_]]
; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]] ; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_CONSTANT i32 15 %2:sgpr(s32) = G_CONSTANT i32 15
%3:sgpr(s32) = G_LSHR %1, %2 %3:sgpr(s32) = G_LSHR %1, %2
%4:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %0, %3 %4:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %0, %3
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
Show All 10 Lines bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_hh_wrong_shift_amt ; GFX9PLUS-LABEL: name: test_build_vector_trunc_s_pack_hh_wrong_shift_amt
; GFX9PLUS: liveins: $sgpr0, $sgpr1 ; GFX9PLUS: liveins: $sgpr0, $sgpr1
; GFX9PLUS-NEXT: {{ $}} ; GFX9PLUS-NEXT: {{ $}}
; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9PLUS-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9PLUS-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 15 ; GFX9PLUS-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 15
; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[S_MOV_B32_]], implicit-def $scc ; GFX9PLUS-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9PLUS-NEXT: [[S_LSHR_B32_1:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GFX9PLUS-NEXT: [[S_LSHR_B32_1:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX9PLUS-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[S_LSHR_B32_]], [[S_LSHR_B32_1]] ; GFX9PLUS-NEXT: [[S_PACK_LL_B32_B16_:%[0-9]+]]:sreg_32 = S_PACK_LL_B32_B16 [[S_LSHR_B32_]], [[S_LSHR_B32_1]]
; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]] ; GFX9PLUS-NEXT: S_ENDPGM 0, implicit [[S_PACK_LL_B32_B16_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_CONSTANT i32 15 %2:sgpr(s32) = G_CONSTANT i32 15
%3:sgpr(s32) = G_LSHR %0, %2 %3:sgpr(s32) = G_LSHR %0, %2
%4:sgpr(s32) = G_LSHR %1, %2 %4:sgpr(s32) = G_LSHR %1, %2
%5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %3, %4 %5:sgpr(<2 x s16>) = G_BUILD_VECTOR_TRUNC %3, %4
▲ Show 20 Lines • Show All 276 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-ctpop.mir

Show All 9 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0 liveins: $sgpr0
; CHECK-LABEL: name: ctpop_s32_ss ; CHECK-LABEL: name: ctpop_s32_ss
; CHECK: liveins: $sgpr0 ; CHECK: liveins: $sgpr0
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[S_BCNT1_I32_B32_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B32 [[COPY]], implicit-def $scc ; CHECK-NEXT: [[S_BCNT1_I32_B32_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B32 [[COPY]], implicit-def dead $scc
; CHECK-NEXT: S_ENDPGM 0, implicit [[S_BCNT1_I32_B32_]] ; CHECK-NEXT: S_ENDPGM 0, implicit [[S_BCNT1_I32_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = G_CTPOP %0 %1:sgpr(s32) = G_CTPOP %0
S_ENDPGM 0, implicit %1 S_ENDPGM 0, implicit %1
... ...
--- ---
name: ctpop_s32_vs name: ctpop_s32_vs
▲ Show 20 Lines • Show All 98 Lines • ▼ Show 20 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; CHECK-LABEL: name: add_ctpop_s32_s_ss_commute0 ; CHECK-LABEL: name: add_ctpop_s32_s_ss_commute0
; CHECK: liveins: $sgpr0, $sgpr1 ; CHECK: liveins: $sgpr0, $sgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[S_BCNT1_I32_B32_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B32 [[COPY]], implicit-def $scc ; CHECK-NEXT: [[S_BCNT1_I32_B32_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B32 [[COPY]], implicit-def dead $scc
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BCNT1_I32_B32_]], [[COPY1]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_BCNT1_I32_B32_]], [[COPY1]], implicit-def dead $scc
; CHECK-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]] ; CHECK-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_CTPOP %0 %2:sgpr(s32) = G_CTPOP %0
%3:sgpr(s32) = G_ADD %2, %1 %3:sgpr(s32) = G_ADD %2, %1
S_ENDPGM 0, implicit %3 S_ENDPGM 0, implicit %3
... ...
▲ Show 20 Lines • Show All 59 Lines • ▼ Show 20 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; CHECK-LABEL: name: add_ctpop_s32_s_sv_commute0 ; CHECK-LABEL: name: add_ctpop_s32_s_sv_commute0
; CHECK: liveins: $sgpr0, $vgpr0 ; CHECK: liveins: $sgpr0, $vgpr0
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[S_BCNT1_I32_B32_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B32 [[COPY]], implicit-def $scc ; CHECK-NEXT: [[S_BCNT1_I32_B32_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B32 [[COPY]], implicit-def dead $scc
; CHECK-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[S_BCNT1_I32_B32_]], [[COPY1]], 0, implicit $exec ; CHECK-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[S_BCNT1_I32_B32_]], [[COPY1]], 0, implicit $exec
; CHECK-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_]] ; CHECK-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%2:sgpr(s32) = G_CTPOP %0 %2:sgpr(s32) = G_CTPOP %0
%3:vgpr(s32) = G_ADD %2, %1 %3:vgpr(s32) = G_ADD %2, %1
S_ENDPGM 0, implicit %3 S_ENDPGM 0, implicit %3
... ...
--- ---
name: ctpop_s64_ss name: ctpop_s64_ss
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1 liveins: $sgpr0_sgpr1
; CHECK-LABEL: name: ctpop_s64_ss ; CHECK-LABEL: name: ctpop_s64_ss
; CHECK: liveins: $sgpr0_sgpr1 ; CHECK: liveins: $sgpr0_sgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; CHECK-NEXT: [[S_BCNT1_I32_B64_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B64 [[COPY]], implicit-def $scc ; CHECK-NEXT: [[S_BCNT1_I32_B64_:%[0-9]+]]:sreg_32 = S_BCNT1_I32_B64 [[COPY]], implicit-def dead $scc
; CHECK-NEXT: S_ENDPGM 0, implicit [[S_BCNT1_I32_B64_]] ; CHECK-NEXT: S_ENDPGM 0, implicit [[S_BCNT1_I32_B64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s32) = G_CTPOP %0 %1:sgpr(s32) = G_CTPOP %0
S_ENDPGM 0, implicit %1 S_ENDPGM 0, implicit %1
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-extract-vector-elt.mir

Show All 17 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v2s32 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v2s32
; MOVREL: liveins: $sgpr0_sgpr1, $sgpr2 ; MOVREL: liveins: $sgpr0_sgpr1, $sgpr2
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v2s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v2s32
; GPRIDX: liveins: $sgpr0_sgpr1, $sgpr2 ; GPRIDX: liveins: $sgpr0_sgpr1, $sgpr2
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v3s32 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v3s32
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v3s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v3s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v4s32 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v4s32
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v4s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v4s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v16s32 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v16s32
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v16s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v16s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v32s32 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v32s32
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v32s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v32s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v2s64 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v2s64
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v2s64 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v2s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v4s64 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v4s64
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v4s64 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v4s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v16s64 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v16s64
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v16s64 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v16s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_1 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_1
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub1, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub1, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_1 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_1
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub1, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub1, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_m1 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_m1
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_m1 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_m1
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]] ; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
%0:sgpr(<8 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 %0:sgpr(<8 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
%1:sgpr(s32) = COPY $sgpr8 %1:sgpr(s32) = COPY $sgpr8
%2:sgpr(s32) = G_CONSTANT i32 -1 %2:sgpr(s32) = G_CONSTANT i32 -1
%3:sgpr(s32) = G_ADD %1, %2 %3:sgpr(s32) = G_ADD %1, %2
%4:sgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3 %4:sgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3
Show All 12 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_7 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_7
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub7, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub7, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_7 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_7
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub7, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub7, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_8 ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_8
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_8 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v8s32_idx_offset_8
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]] ; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
%0:sgpr(<8 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 %0:sgpr(<8 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
%1:sgpr(s32) = COPY $sgpr8 %1:sgpr(s32) = COPY $sgpr8
%2:sgpr(s32) = G_CONSTANT i32 8 %2:sgpr(s32) = G_CONSTANT i32 8
%3:sgpr(s32) = G_ADD %1, %2 %3:sgpr(s32) = G_ADD %1, %2
%4:sgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3 %4:sgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3
Show All 12 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_1 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_1
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub2_sub3, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub2_sub3, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_1 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_1
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub2_sub3, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub2_sub3, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
Show All 17 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_2 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_2
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub4_sub5, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub4_sub5, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_2 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_2
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: $m0 = COPY [[COPY1]] ; GPRIDX-NEXT: $m0 = COPY [[COPY1]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub4_sub5, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub4_sub5, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
Show All 15 Lines bb.0:
liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_m1 ; MOVREL-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_m1
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_m1 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v8s64_idx_offset_m1
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]] ; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]]
; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B64_:%[0-9]+]]:sreg_64 = S_MOVRELS_B64 [[COPY]].sub0_sub1, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B64_]]
%0:sgpr(<8 x s64>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 %0:sgpr(<8 x s64>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
%1:sgpr(s32) = COPY $sgpr8 %1:sgpr(s32) = COPY $sgpr8
%2:sgpr(s32) = G_CONSTANT i32 -1 %2:sgpr(s32) = G_CONSTANT i32 -1
%3:sgpr(s32) = G_ADD %1, %2 %3:sgpr(s32) = G_ADD %1, %2
%4:sgpr(s64) = G_EXTRACT_VECTOR_ELT %0, %3 %4:sgpr(s64) = G_EXTRACT_VECTOR_ELT %0, %3
Show All 12 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v2s32 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v2s32
; MOVREL: liveins: $vgpr0_vgpr1, $sgpr2 ; MOVREL: liveins: $vgpr0_vgpr1, $sgpr2
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v2s32 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v2s32
; GPRIDX: liveins: $vgpr0_vgpr1, $sgpr2 ; GPRIDX: liveins: $vgpr0_vgpr1, $sgpr2
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V2_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V2 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V2_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V2 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V2_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V2_]]
%0:vgpr(<2 x s32>) = COPY $vgpr0_vgpr1 %0:vgpr(<2 x s32>) = COPY $vgpr0_vgpr1
Show All 14 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v3s32 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v3s32
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2, $sgpr3 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2, $sgpr3
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v3s32 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v3s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2, $sgpr3 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2, $sgpr3
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V3_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V3 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V3_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V3 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V3_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V3_]]
%0:vgpr(<3 x s32>) = COPY $vgpr0_vgpr1_vgpr2 %0:vgpr(<3 x s32>) = COPY $vgpr0_vgpr1_vgpr2
Show All 14 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v4s32 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v4s32
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $sgpr4 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v4s32 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v4s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $sgpr4 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V4 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V4 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_]]
%0:vgpr(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3 %0:vgpr(<4 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
Show All 14 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]]
%0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 %0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
Show All 14 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v16s32 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v16s32
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_512 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_512 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v16s32 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v16s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_512 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_512 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V16_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V16 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V16_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V16 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V16_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V16_]]
%0:vgpr(<16 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 %0:vgpr(<16 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
Show All 14 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v32s32 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v32s32
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $sgpr40 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $sgpr40
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_1024 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_1024 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v32s32 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v32s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $sgpr40 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31, $sgpr40
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_1024 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_1024 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V32_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V32 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V32_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V32 [[COPY]], [[COPY1]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V32_]]
%0:vgpr(<32 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31 %0:vgpr(<32 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31
Show All 14 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_1 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_1
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub1, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub1, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_1 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_1
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], 11, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], 11, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]]
%0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 %0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
Show All 14 Lines bb.0:
liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_m1 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_m1
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_m1 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_m1
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[S_ADD_I32_]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[S_ADD_I32_]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]]
%0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 %0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
%1:sgpr(s32) = COPY $sgpr8 %1:sgpr(s32) = COPY $sgpr8
%2:sgpr(s32) = G_CONSTANT i32 -1 %2:sgpr(s32) = G_CONSTANT i32 -1
%3:sgpr(s32) = G_ADD %1, %2 %3:sgpr(s32) = G_ADD %1, %2
%4:vgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3 %4:vgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
Show All 11 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_7 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_7
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: $m0 = COPY [[COPY1]] ; MOVREL-NEXT: $m0 = COPY [[COPY1]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub7, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub7, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_7 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_7
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], 71, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], 71, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]]
%0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 %0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
Show All 14 Lines bb.0:
liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_8 ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_8
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_8 ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v8s32_idx_offset_8
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $sgpr8
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[S_ADD_I32_]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V8 [[COPY]], [[S_ADD_I32_]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V8_]]
%0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 %0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
%1:sgpr(s32) = COPY $sgpr8 %1:sgpr(s32) = COPY $sgpr8
%2:sgpr(s32) = G_CONSTANT i32 8 %2:sgpr(s32) = G_CONSTANT i32 8
%3:sgpr(s32) = G_ADD %1, %2 %3:sgpr(s32) = G_ADD %1, %2
%4:vgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3 %4:vgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %3
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
Show All 11 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_s_s32_v4s32_const_idx ; MOVREL-LABEL: name: extract_vector_elt_s_s32_v4s32_const_idx
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]] ; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]]
; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; MOVREL-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v4s32_const_idx ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v4s32_const_idx
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GPRIDX-NEXT: $m0 = COPY [[S_MOV_B32_]] ; GPRIDX-NEXT: $m0 = COPY [[S_MOV_B32_]]
; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]] ; GPRIDX-NEXT: [[S_MOVRELS_B32_:%[0-9]+]]:sreg_32 = S_MOVRELS_B32 [[COPY]].sub0, implicit $m0, implicit [[COPY]]
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_MOVRELS_B32_]]
Show All 15 Lines bb.0:
; MOVREL-LABEL: name: extract_vector_elt_v_s32_v4s32_const_idx ; MOVREL-LABEL: name: extract_vector_elt_v_s32_v4s32_const_idx
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]] ; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]]
; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]] ; MOVREL-NEXT: [[V_MOVRELS_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOVRELS_B32_e32 [[COPY]].sub0, implicit $m0, implicit $exec, implicit [[COPY]]
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_MOVRELS_B32_e32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v4s32_const_idx ; GPRIDX-LABEL: name: extract_vector_elt_v_s32_v4s32_const_idx
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V4 [[COPY]], [[S_MOV_B32_]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_:%[0-9]+]]:vgpr_32 = V_INDIRECT_REG_READ_GPR_IDX_B32_V4 [[COPY]], [[S_MOV_B32_]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_READ_GPR_IDX_B32_V4_]]
%0:vgpr(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3 %0:vgpr(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
%1:sgpr(s32) = G_CONSTANT i32 0 %1:sgpr(s32) = G_CONSTANT i32 0
%2:vgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %1 %2:vgpr(s32) = G_EXTRACT_VECTOR_ELT %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-insert-vector-elt.mir

Show All 16 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1, $sgpr2, $sgpr3 ; MOVREL: liveins: $sgpr0_sgpr1, $sgpr2, $sgpr3
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V2_:%[0-9]+]]:sreg_64 = S_INDIRECT_REG_WRITE_MOVREL_B32_V2 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V2_:%[0-9]+]]:sreg_64 = S_INDIRECT_REG_WRITE_MOVREL_B32_V2 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V2_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V2_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v2s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v2s32
; GPRIDX: liveins: $sgpr0_sgpr1, $sgpr2, $sgpr3 ; GPRIDX: liveins: $sgpr0_sgpr1, $sgpr2, $sgpr3
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V2_:%[0-9]+]]:sreg_64 = S_INDIRECT_REG_WRITE_MOVREL_B32_V2 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V2_:%[0-9]+]]:sreg_64 = S_INDIRECT_REG_WRITE_MOVREL_B32_V2 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3, $sgpr4 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V3_:%[0-9]+]]:sgpr_96 = S_INDIRECT_REG_WRITE_MOVREL_B32_V3 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V3_:%[0-9]+]]:sgpr_96 = S_INDIRECT_REG_WRITE_MOVREL_B32_V3 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V3_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V3_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v3s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v3s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3, $sgpr4 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2, $sgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_96 = COPY $sgpr0_sgpr1_sgpr2
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V3_:%[0-9]+]]:sgpr_96 = S_INDIRECT_REG_WRITE_MOVREL_B32_V3 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V3_:%[0-9]+]]:sgpr_96 = S_INDIRECT_REG_WRITE_MOVREL_B32_V3 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4, $sgpr5 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4, $sgpr5
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v4s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v4s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4, $sgpr5 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4, $sgpr5
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4, $sgpr5, $sgpr6 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4, $sgpr5, $sgpr6
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_160 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_160 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V5_:%[0-9]+]]:sgpr_160 = S_INDIRECT_REG_WRITE_MOVREL_B32_V5 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V5_:%[0-9]+]]:sgpr_160 = S_INDIRECT_REG_WRITE_MOVREL_B32_V5 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V5_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V5_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v5s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v5s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4, $sgpr5, $sgpr6 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4, $sgpr5, $sgpr6
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_160 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_160 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V5_:%[0-9]+]]:sgpr_160 = S_INDIRECT_REG_WRITE_MOVREL_B32_V5 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V5_:%[0-9]+]]:sgpr_160 = S_INDIRECT_REG_WRITE_MOVREL_B32_V5 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v8s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v8s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8, $sgpr9, $sgpr10 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8, $sgpr9, $sgpr10
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_288 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_288 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V9_:%[0-9]+]]:sgpr_288 = S_INDIRECT_REG_WRITE_MOVREL_B32_V9 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V9_:%[0-9]+]]:sgpr_288 = S_INDIRECT_REG_WRITE_MOVREL_B32_V9 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V9_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V9_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v9s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v9s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8, $sgpr9, $sgpr10 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8, $sgpr9, $sgpr10
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_288 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_288 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V9_:%[0-9]+]]:sgpr_288 = S_INDIRECT_REG_WRITE_MOVREL_B32_V9 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V9_:%[0-9]+]]:sgpr_288 = S_INDIRECT_REG_WRITE_MOVREL_B32_V9 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9, $sgpr10, $sgpr11 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9, $sgpr10, $sgpr11
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_320 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_320 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr10 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr10
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr11 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr11
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V10_:%[0-9]+]]:sgpr_320 = S_INDIRECT_REG_WRITE_MOVREL_B32_V10 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V10_:%[0-9]+]]:sgpr_320 = S_INDIRECT_REG_WRITE_MOVREL_B32_V10 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V10_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V10_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v10s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v10s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9, $sgpr10, $sgpr11 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9, $sgpr10, $sgpr11
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_320 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_320 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr10 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr10
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr11 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr11
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V10_:%[0-9]+]]:sgpr_320 = S_INDIRECT_REG_WRITE_MOVREL_B32_V10 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V10_:%[0-9]+]]:sgpr_320 = S_INDIRECT_REG_WRITE_MOVREL_B32_V10 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10, $sgpr11, $sgpr12 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10, $sgpr11, $sgpr12
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_352 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_352 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr11 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr11
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr12 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr12
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V11_:%[0-9]+]]:sgpr_352 = S_INDIRECT_REG_WRITE_MOVREL_B32_V11 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V11_:%[0-9]+]]:sgpr_352 = S_INDIRECT_REG_WRITE_MOVREL_B32_V11 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V11_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V11_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v11s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v11s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10, $sgpr11, $sgpr12 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10, $sgpr11, $sgpr12
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_352 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_352 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr11 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr11
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr12 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr12
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V11_:%[0-9]+]]:sgpr_352 = S_INDIRECT_REG_WRITE_MOVREL_B32_V11 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V11_:%[0-9]+]]:sgpr_352 = S_INDIRECT_REG_WRITE_MOVREL_B32_V11 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11, $sgpr12, $sgpr13 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11, $sgpr12, $sgpr13
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_384 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_384 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr12 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr12
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr13 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr13
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V12_:%[0-9]+]]:sgpr_384 = S_INDIRECT_REG_WRITE_MOVREL_B32_V12 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V12_:%[0-9]+]]:sgpr_384 = S_INDIRECT_REG_WRITE_MOVREL_B32_V12 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V12_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V12_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v12s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v12s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11, $sgpr12, $sgpr13 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11, $sgpr12, $sgpr13
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_384 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_384 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr12 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr12
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr13 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr13
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V12_:%[0-9]+]]:sgpr_384 = S_INDIRECT_REG_WRITE_MOVREL_B32_V12 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V12_:%[0-9]+]]:sgpr_384 = S_INDIRECT_REG_WRITE_MOVREL_B32_V12 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16, $sgpr17 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16, $sgpr17
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr16 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr16
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr17 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr17
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V16_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B32_V16 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V16_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B32_V16 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V16_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V16_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v16s32 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v16s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16, $sgpr17 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16, $sgpr17
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr16 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr16
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr17 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr17
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V16_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B32_V16 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V16_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B32_V16 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40, $sgpr41 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40, $sgpr41
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr41 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr41
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V32_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B32_V32 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V32_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B32_V32 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V32_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V32_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v32s32 ; GPRIDX-LABEL: name: extract_vector_elt_s_s32_v32s32
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40, $sgpr41 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40, $sgpr41
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr40
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr41 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr41
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V32_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B32_V32 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V32_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B32_V32 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4_sgpr5, $sgpr6 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4_sgpr5, $sgpr6
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr4_sgpr5 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr4_sgpr5
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V2_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B64_V2 [[COPY]], [[COPY1]], 4, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V2_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B64_V2 [[COPY]], [[COPY1]], 4, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V2_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V2_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s64_v2s64 ; GPRIDX-LABEL: name: insert_vector_elt_s_s64_v2s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4_sgpr5, $sgpr6 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4_sgpr5, $sgpr6
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr4_sgpr5 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr4_sgpr5
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V2_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B64_V2 [[COPY]], [[COPY1]], 4, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V2_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B64_V2 [[COPY]], [[COPY1]], 4, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr8_sgpr9 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr8_sgpr9
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V4_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B64_V4 [[COPY]], [[COPY1]], 4, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V4_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B64_V4 [[COPY]], [[COPY1]], 4, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V4_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V4_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s64_v4s64 ; GPRIDX-LABEL: name: insert_vector_elt_s_s64_v4s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr8_sgpr9 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr8_sgpr9
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr10
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V4_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B64_V4 [[COPY]], [[COPY1]], 4, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V4_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B64_V4 [[COPY]], [[COPY1]], 4, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16_sgpr17, $sgpr18 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16_sgpr17, $sgpr18
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr16_sgpr17 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr16_sgpr17
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr18 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr18
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V8_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B64_V8 [[COPY]], [[COPY1]], 4, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V8_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B64_V8 [[COPY]], [[COPY1]], 4, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s64_v8s64 ; GPRIDX-LABEL: name: insert_vector_elt_s_s64_v8s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16_sgpr17, $sgpr18 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15, $sgpr16_sgpr17, $sgpr18
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_512 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr16_sgpr17 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr16_sgpr17
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr18 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr18
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V8_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B64_V8 [[COPY]], [[COPY1]], 4, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V8_:%[0-9]+]]:sgpr_512 = S_INDIRECT_REG_WRITE_MOVREL_B64_V8 [[COPY]], [[COPY1]], 4, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40_sgpr41, $sgpr42 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40_sgpr41, $sgpr42
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr40_sgpr41 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr40_sgpr41
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr42 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr42
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V16_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B64_V16 [[COPY]], [[COPY1]], 4, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V16_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B64_V16 [[COPY]], [[COPY1]], 4, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V16_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B64_V16_]]
;
; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v16s64 ; GPRIDX-LABEL: name: extract_vector_elt_s_s64_v16s64
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40_sgpr41, $sgpr42 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31, $sgpr40_sgpr41, $sgpr42
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_1024 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7_sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15_sgpr16_sgpr17_sgpr18_sgpr19_sgpr20_sgpr21_sgpr22_sgpr23_sgpr24_sgpr25_sgpr26_sgpr27_sgpr28_sgpr29_sgpr30_sgpr31
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr40_sgpr41 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr40_sgpr41
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr42 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr42
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V16_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B64_V16 [[COPY]], [[COPY1]], 4, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B64_V16_:%[0-9]+]]:sgpr_1024 = S_INDIRECT_REG_WRITE_MOVREL_B64_V16 [[COPY]], [[COPY1]], 4, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $vgpr0_vgpr1, $vgpr2, $sgpr3 ; MOVREL: liveins: $vgpr0_vgpr1, $vgpr2, $sgpr3
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V2_:%[0-9]+]]:vreg_64 = V_INDIRECT_REG_WRITE_MOVREL_B32_V2 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V2_:%[0-9]+]]:vreg_64 = V_INDIRECT_REG_WRITE_MOVREL_B32_V2 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V2_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V2_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v2s32 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v2s32
; GPRIDX: liveins: $vgpr0_vgpr1, $vgpr2, $sgpr3 ; GPRIDX: liveins: $vgpr0_vgpr1, $vgpr2, $sgpr3
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V2_:%[0-9]+]]:vreg_64 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V2 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V2_:%[0-9]+]]:vreg_64 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V2 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V2_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V2_]]
Show All 17 Lines bb.0:
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3, $sgpr4 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V3_:%[0-9]+]]:vreg_96 = V_INDIRECT_REG_WRITE_MOVREL_B32_V3 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V3_:%[0-9]+]]:vreg_96 = V_INDIRECT_REG_WRITE_MOVREL_B32_V3 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V3_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V3_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v3s32 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v3s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3, $sgpr4 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2, $vgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V3_:%[0-9]+]]:vreg_96 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V3 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V3_:%[0-9]+]]:vreg_96 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V3 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V3_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V3_]]
Show All 17 Lines bb.0:
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4, $vgpr5 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4, $vgpr5
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v4s32 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v4s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4, $vgpr5 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4, $vgpr5
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_]]
Show All 17 Lines bb.0:
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4, $vgpr5, $sgpr6 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4, $vgpr5, $sgpr6
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_160 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_160 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr5 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr5
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V5_:%[0-9]+]]:vreg_160 = V_INDIRECT_REG_WRITE_MOVREL_B32_V5 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V5_:%[0-9]+]]:vreg_160 = V_INDIRECT_REG_WRITE_MOVREL_B32_V5 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V5_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V5_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v5s32 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v5s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4, $vgpr5, $sgpr6 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4, $vgpr5, $sgpr6
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_160 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_160 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr5 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr5
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V5_:%[0-9]+]]:vreg_160 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V5 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V5_:%[0-9]+]]:vreg_160 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V5 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V5_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V5_]]
Show All 17 Lines bb.0:
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v8s32 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v8s32
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], [[COPY2]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_]]
Show All 17 Lines bb.0:
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 11, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 11, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v8s32_add_1 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v8s32_add_1
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], [[COPY2]], 11, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], [[COPY2]], 11, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_]]
Show All 17 Lines bb.0:
; MOVREL-LABEL: name: insert_vector_elt_vvs_s32_v8s32_add_8 ; MOVREL-LABEL: name: insert_vector_elt_vvs_s32_v8s32_add_8
; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9 ; MOVREL: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v8s32_add_8 ; GPRIDX-LABEL: name: insert_vector_elt_vvs_s32_v8s32_add_8
; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9 ; GPRIDX: liveins: $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7, $vgpr8, $sgpr9
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_256 = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr8
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], [[S_ADD_I32_]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_:%[0-9]+]]:vreg_256 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8 [[COPY]], [[COPY1]], [[S_ADD_I32_]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V8_]]
%0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7 %0:vgpr(<8 x s32>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7
%1:vgpr(s32) = COPY $vgpr8 %1:vgpr(s32) = COPY $vgpr8
%2:sgpr(s32) = COPY $sgpr9 %2:sgpr(s32) = COPY $sgpr9
%3:sgpr(s32) = G_CONSTANT i32 8 %3:sgpr(s32) = G_CONSTANT i32 8
%4:sgpr(s32) = G_ADD %2, %3 %4:sgpr(s32) = G_ADD %2, %3
%5:vgpr(<8 x s32>) = G_INSERT_VECTOR_ELT %0, %1, %4 %5:vgpr(<8 x s32>) = G_INSERT_VECTOR_ELT %0, %1, %4
Show All 13 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: $m0 = COPY [[COPY2]] ; MOVREL-NEXT: $m0 = COPY [[COPY2]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 11, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 11, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v8s32_add_1 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v8s32_add_1
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: $m0 = COPY [[COPY2]] ; GPRIDX-NEXT: $m0 = COPY [[COPY2]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 11, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 11, implicit $m0
Show All 18 Lines bb.0:
; MOVREL-LABEL: name: insert_vector_elt_s_s32_v8s32_add_8 ; MOVREL-LABEL: name: insert_vector_elt_s_s32_v8s32_add_8
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; MOVREL-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def $scc ; MOVREL-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def dead $scc
; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]] ; MOVREL-NEXT: $m0 = COPY [[S_ADD_I32_]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v8s32_add_8 ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v8s32_add_8
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7, $sgpr8, $sgpr9
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_256 = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr8
; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9 ; GPRIDX-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr9
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 8
; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def $scc ; GPRIDX-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY2]], [[S_MOV_B32_]], implicit-def dead $scc
; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]] ; GPRIDX-NEXT: $m0 = COPY [[S_ADD_I32_]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_:%[0-9]+]]:sgpr_256 = S_INDIRECT_REG_WRITE_MOVREL_B32_V8 [[COPY]], [[COPY1]], 3, implicit $m0
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V8_]]
%0:sgpr(<8 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7 %0:sgpr(<8 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
%1:sgpr(s32) = COPY $sgpr8 %1:sgpr(s32) = COPY $sgpr8
%2:sgpr(s32) = COPY $sgpr9 %2:sgpr(s32) = COPY $sgpr9
%3:sgpr(s32) = G_CONSTANT i32 8 %3:sgpr(s32) = G_CONSTANT i32 8
%4:sgpr(s32) = G_ADD %2, %3 %4:sgpr(s32) = G_ADD %2, %3
Show All 16 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]] ; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]]
; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0 ; MOVREL-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0
; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v4s32_const_idx ; GPRIDX-LABEL: name: insert_vector_elt_s_s32_v4s32_const_idx
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GPRIDX-NEXT: $m0 = COPY [[S_MOV_B32_]] ; GPRIDX-NEXT: $m0 = COPY [[S_MOV_B32_]]
; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0 ; GPRIDX-NEXT: [[S_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:sgpr_128 = S_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0
Show All 18 Lines bb.0:
; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; MOVREL: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; MOVREL-NEXT: {{ $}} ; MOVREL-NEXT: {{ $}}
; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; MOVREL-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; MOVREL-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; MOVREL-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]] ; MOVREL-NEXT: $m0 = COPY [[S_MOV_B32_]]
; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec ; MOVREL-NEXT: [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_MOVREL_B32_V4 [[COPY]], [[COPY1]], 3, implicit $m0, implicit $exec
; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_]] ; MOVREL-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_MOVREL_B32_V4_]]
;
; GPRIDX-LABEL: name: insert_vector_elt_v_s32_v4s32_const_idx ; GPRIDX-LABEL: name: insert_vector_elt_v_s32_v4s32_const_idx
; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4 ; GPRIDX: liveins: $sgpr0_sgpr1_sgpr2_sgpr3, $sgpr4
; GPRIDX-NEXT: {{ $}} ; GPRIDX-NEXT: {{ $}}
; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3 ; GPRIDX-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $sgpr0_sgpr1_sgpr2_sgpr3
; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GPRIDX-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GPRIDX-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4 [[COPY]], [[COPY1]], [[S_MOV_B32_]], 3, implicit-def $m0, implicit $m0, implicit $exec ; GPRIDX-NEXT: [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_:%[0-9]+]]:vreg_128 = V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4 [[COPY]], [[COPY1]], [[S_MOV_B32_]], 3, implicit-def $m0, implicit $m0, implicit $exec
; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_]] ; GPRIDX-NEXT: S_ENDPGM 0, implicit [[V_INDIRECT_REG_WRITE_GPR_IDX_B32_V4_]]
%0:vgpr(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3 %0:vgpr(<4 x s32>) = COPY $sgpr0_sgpr1_sgpr2_sgpr3
%1:sgpr(s32) = COPY $sgpr4 %1:sgpr(s32) = COPY $sgpr4
%2:sgpr(s32) = G_CONSTANT i32 0 %2:sgpr(s32) = G_CONSTANT i32 0
%3:vgpr(<4 x s32>) = G_INSERT_VECTOR_ELT %0, %1, %2 %3:vgpr(<4 x s32>) = G_INSERT_VECTOR_ELT %0, %1, %2
S_ENDPGM 0, implicit %3 S_ENDPGM 0, implicit %3
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-lshr.mir

Show All 13 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX6-LABEL: name: lshr_s32_ss ; GFX6-LABEL: name: lshr_s32_ss
; GFX6: liveins: $sgpr0, $sgpr1 ; GFX6: liveins: $sgpr0, $sgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]]
;
; GFX7-LABEL: name: lshr_s32_ss ; GFX7-LABEL: name: lshr_s32_ss
; GFX7: liveins: $sgpr0, $sgpr1 ; GFX7: liveins: $sgpr0, $sgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX7-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX7-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]]
;
; GFX8-LABEL: name: lshr_s32_ss ; GFX8-LABEL: name: lshr_s32_ss
; GFX8: liveins: $sgpr0, $sgpr1 ; GFX8: liveins: $sgpr0, $sgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]]
;
; GFX9-LABEL: name: lshr_s32_ss ; GFX9-LABEL: name: lshr_s32_ss
; GFX9: liveins: $sgpr0, $sgpr1 ; GFX9: liveins: $sgpr0, $sgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]]
;
; GFX10-LABEL: name: lshr_s32_ss ; GFX10-LABEL: name: lshr_s32_ss
; GFX10: liveins: $sgpr0, $sgpr1 ; GFX10: liveins: $sgpr0, $sgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_LSHR_B32_:%[0-9]+]]:sreg_32 = S_LSHR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_LSHR %0, %1 %2:sgpr(s32) = G_LSHR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: lshr_s32_sv name: lshr_s32_sv
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; GFX6-LABEL: name: lshr_s32_sv ; GFX6-LABEL: name: lshr_s32_sv
; GFX6: liveins: $sgpr0, $vgpr0 ; GFX6: liveins: $sgpr0, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX7-LABEL: name: lshr_s32_sv ; GFX7-LABEL: name: lshr_s32_sv
; GFX7: liveins: $sgpr0, $vgpr0 ; GFX7: liveins: $sgpr0, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX8-LABEL: name: lshr_s32_sv ; GFX8-LABEL: name: lshr_s32_sv
; GFX8: liveins: $sgpr0, $vgpr0 ; GFX8: liveins: $sgpr0, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX9-LABEL: name: lshr_s32_sv ; GFX9-LABEL: name: lshr_s32_sv
; GFX9: liveins: $sgpr0, $vgpr0 ; GFX9: liveins: $sgpr0, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX10-LABEL: name: lshr_s32_sv ; GFX10-LABEL: name: lshr_s32_sv
; GFX10: liveins: $sgpr0, $vgpr0 ; GFX10: liveins: $sgpr0, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
Show All 12 Lines bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; GFX6-LABEL: name: lshr_s32_vs ; GFX6-LABEL: name: lshr_s32_vs
; GFX6: liveins: $sgpr0, $vgpr0 ; GFX6: liveins: $sgpr0, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX7-LABEL: name: lshr_s32_vs ; GFX7-LABEL: name: lshr_s32_vs
; GFX7: liveins: $sgpr0, $vgpr0 ; GFX7: liveins: $sgpr0, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX8-LABEL: name: lshr_s32_vs ; GFX8-LABEL: name: lshr_s32_vs
; GFX8: liveins: $sgpr0, $vgpr0 ; GFX8: liveins: $sgpr0, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX9-LABEL: name: lshr_s32_vs ; GFX9-LABEL: name: lshr_s32_vs
; GFX9: liveins: $sgpr0, $vgpr0 ; GFX9: liveins: $sgpr0, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX10-LABEL: name: lshr_s32_vs ; GFX10-LABEL: name: lshr_s32_vs
; GFX10: liveins: $sgpr0, $vgpr0 ; GFX10: liveins: $sgpr0, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 12 Lines bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
; GFX6-LABEL: name: lshr_s32_vv ; GFX6-LABEL: name: lshr_s32_vv
; GFX6: liveins: $vgpr0, $vgpr1 ; GFX6: liveins: $vgpr0, $vgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX6-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX7-LABEL: name: lshr_s32_vv ; GFX7-LABEL: name: lshr_s32_vv
; GFX7: liveins: $vgpr0, $vgpr1 ; GFX7: liveins: $vgpr0, $vgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX8-LABEL: name: lshr_s32_vv ; GFX8-LABEL: name: lshr_s32_vv
; GFX8: liveins: $vgpr0, $vgpr1 ; GFX8: liveins: $vgpr0, $vgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX9-LABEL: name: lshr_s32_vv ; GFX9-LABEL: name: lshr_s32_vv
; GFX9: liveins: $vgpr0, $vgpr1 ; GFX9: liveins: $vgpr0, $vgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
;
; GFX10-LABEL: name: lshr_s32_vv ; GFX10-LABEL: name: lshr_s32_vv
; GFX10: liveins: $vgpr0, $vgpr1 ; GFX10: liveins: $vgpr0, $vgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX10-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $sgpr2 liveins: $sgpr0_sgpr1, $sgpr2
; GFX6-LABEL: name: lshr_s64_ss ; GFX6-LABEL: name: lshr_s64_ss
; GFX6: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX6: liveins: $sgpr0_sgpr1, $sgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]]
;
; GFX7-LABEL: name: lshr_s64_ss ; GFX7-LABEL: name: lshr_s64_ss
; GFX7: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX7: liveins: $sgpr0_sgpr1, $sgpr2
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX7-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]]
;
; GFX8-LABEL: name: lshr_s64_ss ; GFX8-LABEL: name: lshr_s64_ss
; GFX8: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0_sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]]
;
; GFX9-LABEL: name: lshr_s64_ss ; GFX9-LABEL: name: lshr_s64_ss
; GFX9: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0_sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]]
;
; GFX10-LABEL: name: lshr_s64_ss ; GFX10-LABEL: name: lshr_s64_ss
; GFX10: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX10: liveins: $sgpr0_sgpr1, $sgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX10-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_LSHR_B64_:%[0-9]+]]:sreg_64 = S_LSHR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHR_B64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s32) = COPY $sgpr2 %1:sgpr(s32) = COPY $sgpr2
%2:sgpr(s64) = G_LSHR %0, %1 %2:sgpr(s64) = G_LSHR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: lshr_s64_sv name: lshr_s64_sv
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $vgpr0 liveins: $sgpr0_sgpr1, $vgpr0
; GFX6-LABEL: name: lshr_s64_sv ; GFX6-LABEL: name: lshr_s64_sv
; GFX6: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX6: liveins: $sgpr0_sgpr1, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]]
;
; GFX7-LABEL: name: lshr_s64_sv ; GFX7-LABEL: name: lshr_s64_sv
; GFX7: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX7: liveins: $sgpr0_sgpr1, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]]
;
; GFX8-LABEL: name: lshr_s64_sv ; GFX8-LABEL: name: lshr_s64_sv
; GFX8: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0_sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
;
; GFX9-LABEL: name: lshr_s64_sv ; GFX9-LABEL: name: lshr_s64_sv
; GFX9: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0_sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
;
; GFX10-LABEL: name: lshr_s64_sv ; GFX10-LABEL: name: lshr_s64_sv
; GFX10: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX10: liveins: $sgpr0_sgpr1, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
Show All 12 Lines bb.0:
liveins: $sgpr0, $vgpr0_vgpr1 liveins: $sgpr0, $vgpr0_vgpr1
; GFX6-LABEL: name: lshr_s64_vs ; GFX6-LABEL: name: lshr_s64_vs
; GFX6: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX6: liveins: $sgpr0, $vgpr0_vgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]]
;
; GFX7-LABEL: name: lshr_s64_vs ; GFX7-LABEL: name: lshr_s64_vs
; GFX7: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX7: liveins: $sgpr0, $vgpr0_vgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]]
;
; GFX8-LABEL: name: lshr_s64_vs ; GFX8-LABEL: name: lshr_s64_vs
; GFX8: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX8: liveins: $sgpr0, $vgpr0_vgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
;
; GFX9-LABEL: name: lshr_s64_vs ; GFX9-LABEL: name: lshr_s64_vs
; GFX9: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX9: liveins: $sgpr0, $vgpr0_vgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
;
; GFX10-LABEL: name: lshr_s64_vs ; GFX10-LABEL: name: lshr_s64_vs
; GFX10: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX10: liveins: $sgpr0, $vgpr0_vgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
%0:vgpr(s64) = COPY $vgpr0_vgpr1 %0:vgpr(s64) = COPY $vgpr0_vgpr1
Show All 12 Lines bb.0:
liveins: $vgpr0_vgpr1, $vgpr2 liveins: $vgpr0_vgpr1, $vgpr2
; GFX6-LABEL: name: lshr_s64_vv ; GFX6-LABEL: name: lshr_s64_vv
; GFX6: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX6-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]]
;
; GFX7-LABEL: name: lshr_s64_vv ; GFX7-LABEL: name: lshr_s64_vv
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX7-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_LSHR_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHR_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHR_B64_e64_]]
;
; GFX8-LABEL: name: lshr_s64_vv ; GFX8-LABEL: name: lshr_s64_vv
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
;
; GFX9-LABEL: name: lshr_s64_vv ; GFX9-LABEL: name: lshr_s64_vv
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
;
; GFX10-LABEL: name: lshr_s64_vv ; GFX10-LABEL: name: lshr_s64_vv
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX10-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHRREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHRREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHRREV_B64_e64_]]
%0:vgpr(s64) = COPY $vgpr0_vgpr1 %0:vgpr(s64) = COPY $vgpr0_vgpr1
%1:vgpr(s32) = COPY $vgpr2 %1:vgpr(s32) = COPY $vgpr2
%2:vgpr(s64) = G_LSHR %0, %1 %2:vgpr(s64) = G_LSHR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-or.mir

Show All 16 Lines bb.0:
; WAVE64-LABEL: name: or_s1_vcc_vcc_vcc ; WAVE64-LABEL: name: or_s1_vcc_vcc_vcc
; WAVE64: liveins: $vgpr0, $vgpr1 ; WAVE64: liveins: $vgpr0, $vgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec ; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
; ;
; WAVE32-LABEL: name: or_s1_vcc_vcc_vcc ; WAVE32-LABEL: name: or_s1_vcc_vcc_vcc
; WAVE32: liveins: $vgpr0, $vgpr1 ; WAVE32: liveins: $vgpr0, $vgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec ; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(s32) = G_CONSTANT i32 0 %2:vgpr(s32) = G_CONSTANT i32 0
%3:vcc(s1) = G_ICMP intpred(eq), %0, %2 %3:vcc(s1) = G_ICMP intpred(eq), %0, %2
%4:vcc(s1) = G_ICMP intpred(eq), %1, %2 %4:vcc(s1) = G_ICMP intpred(eq), %1, %2
%5:vcc(s1) = G_OR %3, %4 %5:vcc(s1) = G_OR %3, %4
S_ENDPGM 0, implicit %5 S_ENDPGM 0, implicit %5
▲ Show 20 Lines • Show All 109 Lines • ▼ Show 20 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; WAVE64-LABEL: name: or_s32_sgpr_sgpr_sgpr ; WAVE64-LABEL: name: or_s32_sgpr_sgpr_sgpr
; WAVE64: liveins: $sgpr0, $sgpr1 ; WAVE64: liveins: $sgpr0, $sgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE64-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
; ;
; WAVE32-LABEL: name: or_s32_sgpr_sgpr_sgpr ; WAVE32-LABEL: name: or_s32_sgpr_sgpr_sgpr
; WAVE32: liveins: $sgpr0, $sgpr1 ; WAVE32: liveins: $sgpr0, $sgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_OR %0, %1 %2:sgpr(s32) = G_OR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: or_s64_sgpr_sgpr_sgpr name: or_s64_sgpr_sgpr_sgpr
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE64-LABEL: name: or_s64_sgpr_sgpr_sgpr ; WAVE64-LABEL: name: or_s64_sgpr_sgpr_sgpr
; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 ; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
; ;
; WAVE32-LABEL: name: or_s64_sgpr_sgpr_sgpr ; WAVE32-LABEL: name: or_s64_sgpr_sgpr_sgpr
; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 ; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
; WAVE32-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s64) = COPY $sgpr2_sgpr3 %1:sgpr(s64) = COPY $sgpr2_sgpr3
%2:sgpr(s64) = G_OR %0, %1 %2:sgpr(s64) = G_OR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 198 Lines • ▼ Show 20 Lines bb.0:
; WAVE64: liveins: $vgpr0, $vgpr1 ; WAVE64: liveins: $vgpr0, $vgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
; ;
; WAVE32-LABEL: name: or_s1_vcc_copy_to_vcc ; WAVE32-LABEL: name: or_s1_vcc_copy_to_vcc
; WAVE32: liveins: $vgpr0, $vgpr1 ; WAVE32: liveins: $vgpr0, $vgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(s1) = G_TRUNC %0 %2:vgpr(s1) = G_TRUNC %0
%3:vgpr(s1) = G_TRUNC %1 %3:vgpr(s1) = G_TRUNC %1
%4:vcc(s1) = COPY %2 %4:vcc(s1) = COPY %2
%5:vcc(s1) = COPY %3 %5:vcc(s1) = COPY %3
%6:vcc(s1) = G_OR %4, %5 %6:vcc(s1) = G_OR %4, %5
Show All 17 Lines bb.0:
; WAVE64: liveins: $vgpr0, $sgpr0 ; WAVE64: liveins: $vgpr0, $sgpr0
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B64_]] ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B64_]]
; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]]
; ;
; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave32 ; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
; WAVE32: liveins: $vgpr0, $sgpr0 ; WAVE32: liveins: $vgpr0, $sgpr0
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B32_]] ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B32_]]
; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]]
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%0:vgpr(s1) = G_TRUNC %1(s32) %0:vgpr(s1) = G_TRUNC %1(s32)
%sgpr0:sgpr(s32) = COPY $sgpr0 %sgpr0:sgpr(s32) = COPY $sgpr0
%2:sgpr(s1) = G_TRUNC %sgpr0 %2:sgpr(s1) = G_TRUNC %sgpr0
%6:sgpr(s32) = G_CONSTANT i32 0 %6:sgpr(s32) = G_CONSTANT i32 0
%7:sgpr(p1) = G_IMPLICIT_DEF %7:sgpr(p1) = G_IMPLICIT_DEF
Show All 19 Lines bb.0:
; WAVE64: liveins: $vgpr0, $sgpr0 ; WAVE64: liveins: $vgpr0, $sgpr0
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
; ;
; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave64 ; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
; WAVE32: liveins: $vgpr0, $sgpr0 ; WAVE32: liveins: $vgpr0, $sgpr0
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_OR_B32_]] ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_OR_B32_]]
; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]]
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%0:vgpr(s1) = G_TRUNC %1(s32) %0:vgpr(s1) = G_TRUNC %1(s32)
%sgpr0:sgpr(s32) = COPY $sgpr0 %sgpr0:sgpr(s32) = COPY $sgpr0
%2:sgpr(s1) = G_TRUNC %sgpr0 %2:sgpr(s1) = G_TRUNC %sgpr0
%6:sgpr(s32) = G_CONSTANT i32 0 %6:sgpr(s32) = G_CONSTANT i32 0
%7:sgpr(p1) = G_IMPLICIT_DEF %7:sgpr(p1) = G_IMPLICIT_DEF
%9:vcc(s1) = COPY %0(s1) %9:vcc(s1) = COPY %0(s1)
%10:vcc(s1) = COPY %2(s1) %10:vcc(s1) = COPY %2(s1)
%8:vcc(s1) = G_OR %9, %10 %8:vcc(s1) = G_OR %9, %10
%3:sreg_64_xexec(s1) = COPY %8(s1) %3:sreg_64_xexec(s1) = COPY %8(s1)
S_ENDPGM 0, implicit %3 S_ENDPGM 0, implicit %3
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-pattern-add3.mir

Show All 14 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1, $sgpr2 liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-LABEL: name: add_s32_sgpr_sgpr_sgpr ; GFX8-LABEL: name: add_s32_sgpr_sgpr_sgpr
; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_ADD_I32_]], [[COPY2]], implicit-def $scc ; GFX8-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_ADD_I32_]], [[COPY2]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_1]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_1]]
;
; GFX9-LABEL: name: add_s32_sgpr_sgpr_sgpr ; GFX9-LABEL: name: add_s32_sgpr_sgpr_sgpr
; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_ADD_I32_]], [[COPY2]], implicit-def $scc ; GFX9-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[S_ADD_I32_]], [[COPY2]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_1]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_ADD_I32_1]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = COPY $sgpr2 %2:sgpr(s32) = COPY $sgpr2
%3:sgpr(s32) = G_ADD %0, %1 %3:sgpr(s32) = G_ADD %0, %1
%4:sgpr(s32) = G_ADD %3, %2 %4:sgpr(s32) = G_ADD %3, %2
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
... ...
Show All 12 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]]
;
; GFX9-LABEL: name: add_s32_vgpr_vgpr_vgpr ; GFX9-LABEL: name: add_s32_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ADD3_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD3_U32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_ADD3_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD3_U32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD3_U32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD3_U32_e64_]]
Show All 19 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]], implicit [[V_ADD_CO_U32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]], implicit [[V_ADD_CO_U32_e64_]]
;
; GFX9-LABEL: name: add_s32_vgpr_vgpr_vgpr_multi_use ; GFX9-LABEL: name: add_s32_vgpr_vgpr_vgpr_multi_use
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_ADD_U32_e64_]], [[COPY2]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_ADD_U32_e64_]], [[COPY2]], 0, implicit $exec
Show All 21 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]]
;
; GFX9-LABEL: name: add_p3_vgpr_vgpr_vgpr ; GFX9-LABEL: name: add_p3_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ADD3_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD3_U32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_ADD3_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD3_U32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD3_U32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD3_U32_e64_]]
Show All 20 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_ADD_CO_U32_e64_]], [[COPY2]], 0, implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]]
;
; GFX9-LABEL: name: add_p5_vgpr_vgpr_vgpr ; GFX9-LABEL: name: add_p5_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ADD3_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD3_U32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_ADD3_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD3_U32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD3_U32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD3_U32_e64_]]
Show All 20 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY2]], [[V_ADD_CO_U32_e64_]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY2]], [[V_ADD_CO_U32_e64_]], 0, implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]]
;
; GFX9-LABEL: name: add_p3_s32_vgpr_vgpr_vgpr ; GFX9-LABEL: name: add_p3_s32_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY2]], [[V_ADD_U32_e64_]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY2]], [[V_ADD_U32_e64_]], 0, implicit $exec
Show All 21 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY2]], [[V_ADD_CO_U32_e64_]], 0, implicit $exec ; GFX8-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[COPY2]], [[V_ADD_CO_U32_e64_]], 0, implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_ADD_CO_U32_e64_2]]
;
; GFX9-LABEL: name: add_p5_s32_vgpr_vgpr_vgpr ; GFX9-LABEL: name: add_p5_s32_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY]], [[COPY1]], 0, implicit $exec
; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY2]], [[V_ADD_U32_e64_]], 0, implicit $exec ; GFX9-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[COPY2]], [[V_ADD_U32_e64_]], 0, implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD_U32_e64_1]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_ADD_U32_e64_1]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(p5) = COPY $vgpr2 %2:vgpr(p5) = COPY $vgpr2
%3:vgpr(s32) = G_ADD %0, %1 %3:vgpr(s32) = G_ADD %0, %1
%4:vgpr(p5) = G_PTR_ADD %2, %3 %4:vgpr(p5) = G_PTR_ADD %2, %3
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-pattern-and-or.mir

Show All 14 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1, $sgpr2 liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-LABEL: name: and_or_s32_sgpr_sgpr_sgpr ; GFX8-LABEL: name: and_or_s32_sgpr_sgpr_sgpr
; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[COPY2]], implicit-def $scc ; GFX8-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[COPY2]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
;
; GFX9-LABEL: name: and_or_s32_sgpr_sgpr_sgpr ; GFX9-LABEL: name: and_or_s32_sgpr_sgpr_sgpr
; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[COPY2]], implicit-def $scc ; GFX9-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[COPY2]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = COPY $sgpr2 %2:sgpr(s32) = COPY $sgpr2
%3:sgpr(s32) = G_AND %0, %1 %3:sgpr(s32) = G_AND %0, %1
%4:sgpr(s32) = G_OR %3, %2 %4:sgpr(s32) = G_OR %3, %2
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
... ...
Show All 12 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_AND_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_AND_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
;
; GFX9-LABEL: name: and_or_s32_vgpr_vgpr_vgpr ; GFX9-LABEL: name: and_or_s32_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_AND_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_OR_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_AND_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_OR_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_AND_OR_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_AND_OR_B32_e64_]]
Show All 19 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_AND_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_AND_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY2]], [[V_AND_B32_e64_]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY2]], [[V_AND_B32_e64_]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
;
; GFX9-LABEL: name: and_or_s32_vgpr_vgpr_vgpr_commute ; GFX9-LABEL: name: and_or_s32_vgpr_vgpr_vgpr_commute
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_AND_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_OR_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_AND_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_AND_OR_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_AND_OR_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_AND_OR_B32_e64_]]
Show All 16 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1, $vgpr0 liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-LABEL: name: and_or_s32_sgpr_sgpr_vgpr ; GFX8-LABEL: name: and_or_s32_sgpr_sgpr_vgpr
; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_AND_B32_]] ; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_AND_B32_]]
; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
;
; GFX9-LABEL: name: and_or_s32_sgpr_sgpr_vgpr ; GFX9-LABEL: name: and_or_s32_sgpr_sgpr_vgpr
; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_AND_B32_]] ; GFX9-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_AND_B32_]]
; GFX9-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:vgpr(s32) = COPY $vgpr0 %2:vgpr(s32) = COPY $vgpr0
%3:sgpr(s32) = G_AND %0, %1 %3:sgpr(s32) = G_AND %0, %1
%4:vgpr(s32) = COPY %3 %4:vgpr(s32) = COPY %3
%5:vgpr(s32) = G_OR %4, %2 %5:vgpr(s32) = G_OR %4, %2
S_ENDPGM 0, implicit %5 S_ENDPGM 0, implicit %5
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-pattern-or3.mir

Show All 14 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1, $sgpr2 liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-LABEL: name: or_s32_sgpr_sgpr_sgpr ; GFX8-LABEL: name: or_s32_sgpr_sgpr_sgpr
; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[S_OR_B32_1:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_OR_B32_]], [[COPY2]], implicit-def $scc ; GFX8-NEXT: [[S_OR_B32_1:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_OR_B32_]], [[COPY2]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_1]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_1]]
;
; GFX9-LABEL: name: or_s32_sgpr_sgpr_sgpr ; GFX9-LABEL: name: or_s32_sgpr_sgpr_sgpr
; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[S_OR_B32_1:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_OR_B32_]], [[COPY2]], implicit-def $scc ; GFX9-NEXT: [[S_OR_B32_1:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_OR_B32_]], [[COPY2]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_1]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_1]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = COPY $sgpr2 %2:sgpr(s32) = COPY $sgpr2
%3:sgpr(s32) = G_OR %0, %1 %3:sgpr(s32) = G_OR %0, %1
%4:sgpr(s32) = G_OR %3, %2 %4:sgpr(s32) = G_OR %3, %2
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
... ...
Show All 12 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_OR_B32_e64_]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_OR_B32_e64_]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_1]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_1]]
;
; GFX9-LABEL: name: or_s32_vgpr_vgpr_vgpr ; GFX9-LABEL: name: or_s32_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_OR3_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR3_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_OR3_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR3_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_OR3_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_OR3_B32_e64_]]
Show All 19 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_OR_B32_e64_]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_OR_B32_e64_]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_1]], implicit [[V_OR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_1]], implicit [[V_OR_B32_e64_]]
;
; GFX9-LABEL: name: or_s32_vgpr_vgpr_vgpr_multi_use ; GFX9-LABEL: name: or_s32_vgpr_vgpr_vgpr_multi_use
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX9-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX9-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_OR_B32_e64_]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_OR_B32_e64_]], [[COPY2]], implicit $exec
Show All 9 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-pattern-smed3.mir

Show All 38 Lines bb.0:
liveins: $sgpr0, $sgpr1, $sgpr2 liveins: $sgpr0, $sgpr1, $sgpr2
; GFX6-LABEL: name: smed3_s32_sss ; GFX6-LABEL: name: smed3_s32_sss
; GFX6: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX6: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_MAX_I32_:%[0-9]+]]:sreg_32 = S_MAX_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_MAX_I32_:%[0-9]+]]:sreg_32 = S_MAX_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: [[S_MIN_I32_:%[0-9]+]]:sreg_32 = S_MIN_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_MIN_I32_:%[0-9]+]]:sreg_32 = S_MIN_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: [[S_MAX_I32_1:%[0-9]+]]:sreg_32 = S_MAX_I32 [[S_MIN_I32_]], [[COPY2]], implicit-def $scc ; GFX6-NEXT: [[S_MAX_I32_1:%[0-9]+]]:sreg_32 = S_MAX_I32 [[S_MIN_I32_]], [[COPY2]], implicit-def dead $scc
; GFX6-NEXT: [[S_MIN_I32_1:%[0-9]+]]:sreg_32 = S_MIN_I32 [[S_MAX_I32_]], [[S_MAX_I32_1]], implicit-def $scc ; GFX6-NEXT: [[S_MIN_I32_1:%[0-9]+]]:sreg_32 = S_MIN_I32 [[S_MAX_I32_]], [[S_MAX_I32_1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_MIN_I32_1]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_MIN_I32_1]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = COPY $sgpr2 %2:sgpr(s32) = COPY $sgpr2
%3:sgpr(s32) = G_SMAX %0, %1 %3:sgpr(s32) = G_SMAX %0, %1
%4:sgpr(s32) = G_SMIN %0, %1 %4:sgpr(s32) = G_SMIN %0, %1
%5:sgpr(s32) = G_SMAX %4, %2 %5:sgpr(s32) = G_SMAX %4, %2
%6:sgpr(s32) = G_SMIN %3, %5 %6:sgpr(s32) = G_SMIN %3, %5
▲ Show 20 Lines • Show All 119 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-pattern-umed3.mir

Show All 38 Lines bb.0:
liveins: $sgpr0, $sgpr1, $sgpr2 liveins: $sgpr0, $sgpr1, $sgpr2
; GFX6-LABEL: name: umed3_s32_sss ; GFX6-LABEL: name: umed3_s32_sss
; GFX6: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX6: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_MAX_U32_:%[0-9]+]]:sreg_32 = S_MAX_U32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_MAX_U32_:%[0-9]+]]:sreg_32 = S_MAX_U32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: [[S_MIN_U32_:%[0-9]+]]:sreg_32 = S_MIN_U32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_MIN_U32_:%[0-9]+]]:sreg_32 = S_MIN_U32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: [[S_MAX_U32_1:%[0-9]+]]:sreg_32 = S_MAX_U32 [[S_MIN_U32_]], [[COPY2]], implicit-def $scc ; GFX6-NEXT: [[S_MAX_U32_1:%[0-9]+]]:sreg_32 = S_MAX_U32 [[S_MIN_U32_]], [[COPY2]], implicit-def dead $scc
; GFX6-NEXT: [[S_MIN_U32_1:%[0-9]+]]:sreg_32 = S_MIN_U32 [[S_MAX_U32_]], [[S_MAX_U32_1]], implicit-def $scc ; GFX6-NEXT: [[S_MIN_U32_1:%[0-9]+]]:sreg_32 = S_MIN_U32 [[S_MAX_U32_]], [[S_MAX_U32_1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_MIN_U32_1]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_MIN_U32_1]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = COPY $sgpr2 %2:sgpr(s32) = COPY $sgpr2
%3:sgpr(s32) = G_UMAX %0, %1 %3:sgpr(s32) = G_UMAX %0, %1
%4:sgpr(s32) = G_UMIN %0, %1 %4:sgpr(s32) = G_UMIN %0, %1
%5:sgpr(s32) = G_UMAX %4, %2 %5:sgpr(s32) = G_UMAX %4, %2
%6:sgpr(s32) = G_UMIN %3, %5 %6:sgpr(s32) = G_UMIN %3, %5
▲ Show 20 Lines • Show All 119 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-pattern-xor3.mir

Show All 14 Linesbody: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1, $sgpr2 liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-LABEL: name: xor_s32_sgpr_sgpr_sgpr ; GFX8-LABEL: name: xor_s32_sgpr_sgpr_sgpr
; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[S_XOR_B32_1:%[0-9]+]]:sreg_32 = S_XOR_B32 [[S_XOR_B32_]], [[COPY2]], implicit-def $scc ; GFX8-NEXT: [[S_XOR_B32_1:%[0-9]+]]:sreg_32 = S_XOR_B32 [[S_XOR_B32_]], [[COPY2]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_1]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_1]]
;
; GFX9-LABEL: name: xor_s32_sgpr_sgpr_sgpr ; GFX9-LABEL: name: xor_s32_sgpr_sgpr_sgpr
; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[S_XOR_B32_1:%[0-9]+]]:sreg_32 = S_XOR_B32 [[S_XOR_B32_]], [[COPY2]], implicit-def $scc ; GFX9-NEXT: [[S_XOR_B32_1:%[0-9]+]]:sreg_32 = S_XOR_B32 [[S_XOR_B32_]], [[COPY2]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_1]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_1]]
;
; GFX10-LABEL: name: xor_s32_sgpr_sgpr_sgpr ; GFX10-LABEL: name: xor_s32_sgpr_sgpr_sgpr
; GFX10: liveins: $sgpr0, $sgpr1, $sgpr2 ; GFX10: liveins: $sgpr0, $sgpr1, $sgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX10-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX10-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: [[S_XOR_B32_1:%[0-9]+]]:sreg_32 = S_XOR_B32 [[S_XOR_B32_]], [[COPY2]], implicit-def $scc ; GFX10-NEXT: [[S_XOR_B32_1:%[0-9]+]]:sreg_32 = S_XOR_B32 [[S_XOR_B32_]], [[COPY2]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_1]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_1]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = COPY $sgpr2 %2:sgpr(s32) = COPY $sgpr2
%3:sgpr(s32) = G_XOR %0, %1 %3:sgpr(s32) = G_XOR %0, %1
%4:sgpr(s32) = G_XOR %3, %2 %4:sgpr(s32) = G_XOR %3, %2
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
... ...
Show All 12 Lines bb.0:
; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_XOR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[V_XOR_B32_e64_]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_XOR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[V_XOR_B32_e64_]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_1]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_1]]
;
; GFX9-LABEL: name: xor_s32_vgpr_vgpr_vgpr ; GFX9-LABEL: name: xor_s32_vgpr_vgpr_vgpr
; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX9-NEXT: [[V_XOR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[V_XOR_B32_e64_]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_XOR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[V_XOR_B32_e64_]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_1]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_1]]
;
; GFX10-LABEL: name: xor_s32_vgpr_vgpr_vgpr ; GFX10-LABEL: name: xor_s32_vgpr_vgpr_vgpr
; GFX10: liveins: $vgpr0, $vgpr1, $vgpr2 ; GFX10: liveins: $vgpr0, $vgpr1, $vgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX10-NEXT: [[V_XOR3_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR3_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX10-NEXT: [[V_XOR3_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR3_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR3_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR3_B32_e64_]]
Show All 20 Lines bb.0:
liveins: $sgpr0, $sgpr1, $vgpr0 liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy ; GFX8-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy
; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]] ; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]]
; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
;
; GFX9-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy ; GFX9-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy
; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]] ; GFX9-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]]
; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
;
; GFX10-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy ; GFX10-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy
; GFX10: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX10: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]] ; GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]]
; GFX10-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec ; GFX10-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY3]], [[COPY2]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:vgpr(s32) = COPY $vgpr0 %2:vgpr(s32) = COPY $vgpr0
%3:sgpr(s32) = G_XOR %0, %1 %3:sgpr(s32) = G_XOR %0, %1
%4:vgpr(s32) = COPY %3 %4:vgpr(s32) = COPY %3
Show All 13 Lines bb.0:
liveins: $sgpr0, $sgpr1, $vgpr0 liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy_commute ; GFX8-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy_commute
; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]] ; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]]
; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY2]], [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY2]], [[COPY3]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
;
; GFX9-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy_commute ; GFX9-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy_commute
; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]] ; GFX9-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]]
; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY2]], [[COPY3]], implicit $exec ; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY2]], [[COPY3]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
;
; GFX10-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy_commute ; GFX10-LABEL: name: xor_s32_sgpr_sgpr_vgpr_copy_commute
; GFX10: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX10: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]] ; GFX10-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_XOR_B32_]]
; GFX10-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY2]], [[COPY3]], implicit $exec ; GFX10-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[COPY2]], [[COPY3]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:vgpr(s32) = COPY $vgpr0 %2:vgpr(s32) = COPY $vgpr0
%3:sgpr(s32) = G_XOR %0, %1 %3:sgpr(s32) = G_XOR %0, %1
%4:vgpr(s32) = COPY %3 %4:vgpr(s32) = COPY %3
Show All 13 Lines bb.0:
liveins: $sgpr0, $sgpr1, $vgpr0 liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-LABEL: name: xor_s32_sgpr_sgpr_vgpr ; GFX8-LABEL: name: xor_s32_sgpr_sgpr_vgpr
; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[S_XOR_B32_]], [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[S_XOR_B32_]], [[COPY2]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
;
; GFX9-LABEL: name: xor_s32_sgpr_sgpr_vgpr ; GFX9-LABEL: name: xor_s32_sgpr_sgpr_vgpr
; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[S_XOR_B32_]], [[COPY2]], implicit $exec ; GFX9-NEXT: [[V_XOR_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR_B32_e64 [[S_XOR_B32_]], [[COPY2]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_XOR_B32_e64_]]
;
; GFX10-LABEL: name: xor_s32_sgpr_sgpr_vgpr ; GFX10-LABEL: name: xor_s32_sgpr_sgpr_vgpr
; GFX10: liveins: $sgpr0, $sgpr1, $vgpr0 ; GFX10: liveins: $sgpr0, $sgpr1, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_XOR3_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR3_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec ; GFX10-NEXT: [[V_XOR3_B32_e64_:%[0-9]+]]:vgpr_32 = V_XOR3_B32_e64 [[COPY]], [[COPY1]], [[COPY2]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR3_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_XOR3_B32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:vgpr(s32) = COPY $vgpr0 %2:vgpr(s32) = COPY $vgpr0
%3:sgpr(s32) = G_XOR %0, %1 %3:sgpr(s32) = G_XOR %0, %1
%4:vgpr(s32) = G_XOR %3, %2 %4:vgpr(s32) = G_XOR %3, %2
S_ENDPGM 0, implicit %4 S_ENDPGM 0, implicit %4
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-shl.mir

Show All 13 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GFX6-LABEL: name: shl_s32_ss ; GFX6-LABEL: name: shl_s32_ss
; GFX6: liveins: $sgpr0, $sgpr1 ; GFX6: liveins: $sgpr0, $sgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]]
;
; GFX7-LABEL: name: shl_s32_ss ; GFX7-LABEL: name: shl_s32_ss
; GFX7: liveins: $sgpr0, $sgpr1 ; GFX7: liveins: $sgpr0, $sgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX7-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX7-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]]
;
; GFX8-LABEL: name: shl_s32_ss ; GFX8-LABEL: name: shl_s32_ss
; GFX8: liveins: $sgpr0, $sgpr1 ; GFX8: liveins: $sgpr0, $sgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX8-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]]
;
; GFX9-LABEL: name: shl_s32_ss ; GFX9-LABEL: name: shl_s32_ss
; GFX9: liveins: $sgpr0, $sgpr1 ; GFX9: liveins: $sgpr0, $sgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]]
;
; GFX10-LABEL: name: shl_s32_ss ; GFX10-LABEL: name: shl_s32_ss
; GFX10: liveins: $sgpr0, $sgpr1 ; GFX10: liveins: $sgpr0, $sgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX10-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_SHL %0, %1 %2:sgpr(s32) = G_SHL %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: shl_s32_sv name: shl_s32_sv
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; GFX6-LABEL: name: shl_s32_sv ; GFX6-LABEL: name: shl_s32_sv
; GFX6: liveins: $sgpr0, $vgpr0 ; GFX6: liveins: $sgpr0, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX7-LABEL: name: shl_s32_sv ; GFX7-LABEL: name: shl_s32_sv
; GFX7: liveins: $sgpr0, $vgpr0 ; GFX7: liveins: $sgpr0, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX8-LABEL: name: shl_s32_sv ; GFX8-LABEL: name: shl_s32_sv
; GFX8: liveins: $sgpr0, $vgpr0 ; GFX8: liveins: $sgpr0, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX9-LABEL: name: shl_s32_sv ; GFX9-LABEL: name: shl_s32_sv
; GFX9: liveins: $sgpr0, $vgpr0 ; GFX9: liveins: $sgpr0, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX10-LABEL: name: shl_s32_sv ; GFX10-LABEL: name: shl_s32_sv
; GFX10: liveins: $sgpr0, $vgpr0 ; GFX10: liveins: $sgpr0, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
Show All 12 Lines bb.0:
liveins: $sgpr0, $vgpr0 liveins: $sgpr0, $vgpr0
; GFX6-LABEL: name: shl_s32_vs ; GFX6-LABEL: name: shl_s32_vs
; GFX6: liveins: $sgpr0, $vgpr0 ; GFX6: liveins: $sgpr0, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX7-LABEL: name: shl_s32_vs ; GFX7-LABEL: name: shl_s32_vs
; GFX7: liveins: $sgpr0, $vgpr0 ; GFX7: liveins: $sgpr0, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX8-LABEL: name: shl_s32_vs ; GFX8-LABEL: name: shl_s32_vs
; GFX8: liveins: $sgpr0, $vgpr0 ; GFX8: liveins: $sgpr0, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX9-LABEL: name: shl_s32_vs ; GFX9-LABEL: name: shl_s32_vs
; GFX9: liveins: $sgpr0, $vgpr0 ; GFX9: liveins: $sgpr0, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX10-LABEL: name: shl_s32_vs ; GFX10-LABEL: name: shl_s32_vs
; GFX10: liveins: $sgpr0, $vgpr0 ; GFX10: liveins: $sgpr0, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 12 Lines bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
; GFX6-LABEL: name: shl_s32_vv ; GFX6-LABEL: name: shl_s32_vv
; GFX6: liveins: $vgpr0, $vgpr1 ; GFX6: liveins: $vgpr0, $vgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX6-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX6-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX7-LABEL: name: shl_s32_vv ; GFX7-LABEL: name: shl_s32_vv
; GFX7: liveins: $vgpr0, $vgpr1 ; GFX7: liveins: $vgpr0, $vgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX7-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX8-LABEL: name: shl_s32_vv ; GFX8-LABEL: name: shl_s32_vv
; GFX8: liveins: $vgpr0, $vgpr1 ; GFX8: liveins: $vgpr0, $vgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX9-LABEL: name: shl_s32_vv ; GFX9-LABEL: name: shl_s32_vv
; GFX9: liveins: $vgpr0, $vgpr1 ; GFX9: liveins: $vgpr0, $vgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX9-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
;
; GFX10-LABEL: name: shl_s32_vv ; GFX10-LABEL: name: shl_s32_vv
; GFX10: liveins: $vgpr0, $vgpr1 ; GFX10: liveins: $vgpr0, $vgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX10-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B32_e64_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $sgpr2 liveins: $sgpr0_sgpr1, $sgpr2
; GFX6-LABEL: name: shl_s64_ss ; GFX6-LABEL: name: shl_s64_ss
; GFX6: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX6: liveins: $sgpr0_sgpr1, $sgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]]
;
; GFX7-LABEL: name: shl_s64_ss ; GFX7-LABEL: name: shl_s64_ss
; GFX7: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX7: liveins: $sgpr0_sgpr1, $sgpr2
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX7-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]]
;
; GFX8-LABEL: name: shl_s64_ss ; GFX8-LABEL: name: shl_s64_ss
; GFX8: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX8: liveins: $sgpr0_sgpr1, $sgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX8-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]]
;
; GFX9-LABEL: name: shl_s64_ss ; GFX9-LABEL: name: shl_s64_ss
; GFX9: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX9: liveins: $sgpr0_sgpr1, $sgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX9-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]]
;
; GFX10-LABEL: name: shl_s64_ss ; GFX10-LABEL: name: shl_s64_ss
; GFX10: liveins: $sgpr0_sgpr1, $sgpr2 ; GFX10: liveins: $sgpr0_sgpr1, $sgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX10-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def $scc ; GFX10-NEXT: [[S_LSHL_B64_:%[0-9]+]]:sreg_64 = S_LSHL_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[S_LSHL_B64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s32) = COPY $sgpr2 %1:sgpr(s32) = COPY $sgpr2
%2:sgpr(s64) = G_SHL %0, %1 %2:sgpr(s64) = G_SHL %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: shl_s64_sv name: shl_s64_sv
legalized: true legalized: true
regBankSelected: true regBankSelected: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $vgpr0 liveins: $sgpr0_sgpr1, $vgpr0
; GFX6-LABEL: name: shl_s64_sv ; GFX6-LABEL: name: shl_s64_sv
; GFX6: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX6: liveins: $sgpr0_sgpr1, $vgpr0
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]]
;
; GFX7-LABEL: name: shl_s64_sv ; GFX7-LABEL: name: shl_s64_sv
; GFX7: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX7: liveins: $sgpr0_sgpr1, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]]
;
; GFX8-LABEL: name: shl_s64_sv ; GFX8-LABEL: name: shl_s64_sv
; GFX8: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX8: liveins: $sgpr0_sgpr1, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
;
; GFX9-LABEL: name: shl_s64_sv ; GFX9-LABEL: name: shl_s64_sv
; GFX9: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX9: liveins: $sgpr0_sgpr1, $vgpr0
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
;
; GFX10-LABEL: name: shl_s64_sv ; GFX10-LABEL: name: shl_s64_sv
; GFX10: liveins: $sgpr0_sgpr1, $vgpr0 ; GFX10: liveins: $sgpr0_sgpr1, $vgpr0
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
Show All 12 Lines bb.0:
liveins: $sgpr0, $vgpr0_vgpr1 liveins: $sgpr0, $vgpr0_vgpr1
; GFX6-LABEL: name: shl_s64_vs ; GFX6-LABEL: name: shl_s64_vs
; GFX6: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX6: liveins: $sgpr0, $vgpr0_vgpr1
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]]
;
; GFX7-LABEL: name: shl_s64_vs ; GFX7-LABEL: name: shl_s64_vs
; GFX7: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX7: liveins: $sgpr0, $vgpr0_vgpr1
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX7-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]]
;
; GFX8-LABEL: name: shl_s64_vs ; GFX8-LABEL: name: shl_s64_vs
; GFX8: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX8: liveins: $sgpr0, $vgpr0_vgpr1
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX8-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
;
; GFX9-LABEL: name: shl_s64_vs ; GFX9-LABEL: name: shl_s64_vs
; GFX9: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX9: liveins: $sgpr0, $vgpr0_vgpr1
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
;
; GFX10-LABEL: name: shl_s64_vs ; GFX10-LABEL: name: shl_s64_vs
; GFX10: liveins: $sgpr0, $vgpr0_vgpr1 ; GFX10: liveins: $sgpr0, $vgpr0_vgpr1
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX10-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
%0:vgpr(s64) = COPY $vgpr0_vgpr1 %0:vgpr(s64) = COPY $vgpr0_vgpr1
Show All 12 Lines bb.0:
liveins: $vgpr0_vgpr1, $vgpr2 liveins: $vgpr0_vgpr1, $vgpr2
; GFX6-LABEL: name: shl_s64_vv ; GFX6-LABEL: name: shl_s64_vv
; GFX6: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX6: liveins: $vgpr0_vgpr1, $vgpr2
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX6-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]]
;
; GFX7-LABEL: name: shl_s64_vv ; GFX7-LABEL: name: shl_s64_vv
; GFX7: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX7: liveins: $vgpr0_vgpr1, $vgpr2
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX7-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_LSHL_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHL_B64_e64 [[COPY]], [[COPY1]], implicit $exec
; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]] ; GFX7-NEXT: S_ENDPGM 0, implicit [[V_LSHL_B64_e64_]]
;
; GFX8-LABEL: name: shl_s64_vv ; GFX8-LABEL: name: shl_s64_vv
; GFX8: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX8: liveins: $vgpr0_vgpr1, $vgpr2
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX8-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX8-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
;
; GFX9-LABEL: name: shl_s64_vv ; GFX9-LABEL: name: shl_s64_vv
; GFX9: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX9: liveins: $vgpr0_vgpr1, $vgpr2
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX9-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX9-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX9-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
;
; GFX10-LABEL: name: shl_s64_vv ; GFX10-LABEL: name: shl_s64_vv
; GFX10: liveins: $vgpr0_vgpr1, $vgpr2 ; GFX10: liveins: $vgpr0_vgpr1, $vgpr2
; GFX10-NEXT: {{ $}} ; GFX10-NEXT: {{ $}}
; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1 ; GFX10-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX10-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX10-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec ; GFX10-NEXT: [[V_LSHLREV_B64_e64_:%[0-9]+]]:vreg_64 = V_LSHLREV_B64_e64 [[COPY1]], [[COPY]], implicit $exec
; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]] ; GFX10-NEXT: S_ENDPGM 0, implicit [[V_LSHLREV_B64_e64_]]
%0:vgpr(s64) = COPY $vgpr0_vgpr1 %0:vgpr(s64) = COPY $vgpr0_vgpr1
%1:vgpr(s32) = COPY $vgpr2 %1:vgpr(s32) = COPY $vgpr2
%2:vgpr(s64) = G_SHL %0, %1 %2:vgpr(s64) = G_SHL %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-smax.mir

Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GCN-LABEL: name: smax_s32_ss ; GCN-LABEL: name: smax_s32_ss
; GCN: liveins: $sgpr0, $sgpr1 ; GCN: liveins: $sgpr0, $sgpr1
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GCN-NEXT: [[S_MAX_I32_:%[0-9]+]]:sreg_32 = S_MAX_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GCN-NEXT: [[S_MAX_I32_:%[0-9]+]]:sreg_32 = S_MAX_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GCN-NEXT: S_ENDPGM 0, implicit [[S_MAX_I32_]] ; GCN-NEXT: S_ENDPGM 0, implicit [[S_MAX_I32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_SMAX %0, %1 %2:sgpr(s32) = G_SMAX %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 61 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-smin.mir

Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GCN-LABEL: name: smin_s32_ss ; GCN-LABEL: name: smin_s32_ss
; GCN: liveins: $sgpr0, $sgpr1 ; GCN: liveins: $sgpr0, $sgpr1
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GCN-NEXT: [[S_MIN_I32_:%[0-9]+]]:sreg_32 = S_MIN_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GCN-NEXT: [[S_MIN_I32_:%[0-9]+]]:sreg_32 = S_MIN_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GCN-NEXT: S_ENDPGM 0, implicit [[S_MIN_I32_]] ; GCN-NEXT: S_ENDPGM 0, implicit [[S_MIN_I32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_SMIN %0, %1 %2:sgpr(s32) = G_SMIN %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 61 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-sub.mir

Show All 16 Lines bb.0:
; GFX6-LABEL: name: sub_s32 ; GFX6-LABEL: name: sub_s32
; GFX6: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4 ; GFX6: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX6-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX6-NEXT: [[V_SUB_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[COPY2]], [[S_SUB_I32_]], 0, implicit $exec ; GFX6-NEXT: [[V_SUB_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[COPY2]], [[S_SUB_I32_]], 0, implicit $exec
; GFX6-NEXT: [[V_SUB_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[S_SUB_I32_]], [[V_SUB_CO_U32_e64_]], 0, implicit $exec ; GFX6-NEXT: [[V_SUB_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_3:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[S_SUB_I32_]], [[V_SUB_CO_U32_e64_]], 0, implicit $exec
; GFX6-NEXT: [[V_SUB_CO_U32_e64_4:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_5:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[V_SUB_CO_U32_e64_2]], [[COPY2]], 0, implicit $exec ; GFX6-NEXT: [[V_SUB_CO_U32_e64_4:%[0-9]+]]:vgpr_32, dead [[V_SUB_CO_U32_e64_5:%[0-9]+]]:sreg_64 = V_SUB_CO_U32_e64 [[V_SUB_CO_U32_e64_2]], [[COPY2]], 0, implicit $exec
; GFX6-NEXT: S_ENDPGM 0, implicit [[V_SUB_CO_U32_e64_4]] ; GFX6-NEXT: S_ENDPGM 0, implicit [[V_SUB_CO_U32_e64_4]]
;
; GFX9-LABEL: name: sub_s32 ; GFX9-LABEL: name: sub_s32
; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4 ; GFX9: liveins: $sgpr0, $sgpr1, $vgpr0, $vgpr3_vgpr4
; GFX9-NEXT: {{ $}} ; GFX9-NEXT: {{ $}}
; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GFX9-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GFX9-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX9-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX9-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], [[COPY1]], implicit-def $scc ; GFX9-NEXT: [[S_SUB_I32_:%[0-9]+]]:sreg_32 = S_SUB_I32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GFX9-NEXT: [[V_SUB_U32_e64_:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[COPY2]], [[S_SUB_I32_]], 0, implicit $exec ; GFX9-NEXT: [[V_SUB_U32_e64_:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[COPY2]], [[S_SUB_I32_]], 0, implicit $exec
; GFX9-NEXT: [[V_SUB_U32_e64_1:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[S_SUB_I32_]], [[V_SUB_U32_e64_]], 0, implicit $exec ; GFX9-NEXT: [[V_SUB_U32_e64_1:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[S_SUB_I32_]], [[V_SUB_U32_e64_]], 0, implicit $exec
; GFX9-NEXT: [[V_SUB_U32_e64_2:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[V_SUB_U32_e64_1]], [[COPY2]], 0, implicit $exec ; GFX9-NEXT: [[V_SUB_U32_e64_2:%[0-9]+]]:vgpr_32 = V_SUB_U32_e64 [[V_SUB_U32_e64_1]], [[COPY2]], 0, implicit $exec
; GFX9-NEXT: S_ENDPGM 0, implicit [[V_SUB_U32_e64_2]] ; GFX9-NEXT: S_ENDPGM 0, implicit [[V_SUB_U32_e64_2]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:vgpr(s32) = COPY $vgpr0 %2:vgpr(s32) = COPY $vgpr0
%3:vgpr(p1) = COPY $vgpr3_vgpr4 %3:vgpr(p1) = COPY $vgpr3_vgpr4
%4:sgpr(s32) = G_CONSTANT i32 1 %4:sgpr(s32) = G_CONSTANT i32 1
%5:sgpr(s32) = G_CONSTANT i32 4096 %5:sgpr(s32) = G_CONSTANT i32 4096
; sub ss
%6:sgpr(s32) = G_SUB %0, %1 %6:sgpr(s32) = G_SUB %0, %1
; sub vs
%7:vgpr(s32) = G_SUB %2, %6 %7:vgpr(s32) = G_SUB %2, %6
; sub sv
%8:vgpr(s32) = G_SUB %6, %7 %8:vgpr(s32) = G_SUB %6, %7
; sub vv
%9:vgpr(s32) = G_SUB %8, %2 %9:vgpr(s32) = G_SUB %8, %2
S_ENDPGM 0, implicit %9 S_ENDPGM 0, implicit %9
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-umax.mir

Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GCN-LABEL: name: umax_s32_ss ; GCN-LABEL: name: umax_s32_ss
; GCN: liveins: $sgpr0, $sgpr1 ; GCN: liveins: $sgpr0, $sgpr1
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GCN-NEXT: [[S_MAX_U32_:%[0-9]+]]:sreg_32 = S_MAX_U32 [[COPY]], [[COPY1]], implicit-def $scc ; GCN-NEXT: [[S_MAX_U32_:%[0-9]+]]:sreg_32 = S_MAX_U32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GCN-NEXT: S_ENDPGM 0, implicit [[S_MAX_U32_]] ; GCN-NEXT: S_ENDPGM 0, implicit [[S_MAX_U32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_UMAX %0, %1 %2:sgpr(s32) = G_UMAX %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 61 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-umin.mir

Show All 10 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; GCN-LABEL: name: umin_s32_ss ; GCN-LABEL: name: umin_s32_ss
; GCN: liveins: $sgpr0, $sgpr1 ; GCN: liveins: $sgpr0, $sgpr1
; GCN-NEXT: {{ $}} ; GCN-NEXT: {{ $}}
; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; GCN-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; GCN-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; GCN-NEXT: [[S_MIN_U32_:%[0-9]+]]:sreg_32 = S_MIN_U32 [[COPY]], [[COPY1]], implicit-def $scc ; GCN-NEXT: [[S_MIN_U32_:%[0-9]+]]:sreg_32 = S_MIN_U32 [[COPY]], [[COPY1]], implicit-def dead $scc
; GCN-NEXT: S_ENDPGM 0, implicit [[S_MIN_U32_]] ; GCN-NEXT: S_ENDPGM 0, implicit [[S_MIN_U32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_UMIN %0, %1 %2:sgpr(s32) = G_UMIN %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 61 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-xor.mir

Show All 16 Lines bb.0:
; WAVE64-LABEL: name: xor_s1_vcc_vcc_vcc ; WAVE64-LABEL: name: xor_s1_vcc_vcc_vcc
; WAVE64: liveins: $vgpr0, $vgpr1 ; WAVE64: liveins: $vgpr0, $vgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec ; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]]
; ;
; WAVE32-LABEL: name: xor_s1_vcc_vcc_vcc ; WAVE32-LABEL: name: xor_s1_vcc_vcc_vcc
; WAVE32: liveins: $vgpr0, $vgpr1 ; WAVE32: liveins: $vgpr0, $vgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec ; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(s32) = G_CONSTANT i32 0 %2:vgpr(s32) = G_CONSTANT i32 0
%3:vcc(s1) = G_ICMP intpred(eq), %0, %2 %3:vcc(s1) = G_ICMP intpred(eq), %0, %2
%4:vcc(s1) = G_ICMP intpred(eq), %1, %2 %4:vcc(s1) = G_ICMP intpred(eq), %1, %2
%5:vcc(s1) = G_XOR %3, %4 %5:vcc(s1) = G_XOR %3, %4
S_ENDPGM 0, implicit %5 S_ENDPGM 0, implicit %5
▲ Show 20 Lines • Show All 110 Lines • ▼ Show 20 Lines
body: | body: |
bb.0: bb.0:
liveins: $sgpr0, $sgpr1 liveins: $sgpr0, $sgpr1
; WAVE64-LABEL: name: xor_s32_sgpr_sgpr_sgpr ; WAVE64-LABEL: name: xor_s32_sgpr_sgpr_sgpr
; WAVE64: liveins: $sgpr0, $sgpr1 ; WAVE64: liveins: $sgpr0, $sgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE64-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]]
; ;
; WAVE32-LABEL: name: xor_s32_sgpr_sgpr_sgpr ; WAVE32-LABEL: name: xor_s32_sgpr_sgpr_sgpr
; WAVE32: liveins: $sgpr0, $sgpr1 ; WAVE32: liveins: $sgpr0, $sgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32 = S_XOR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]]
%0:sgpr(s32) = COPY $sgpr0 %0:sgpr(s32) = COPY $sgpr0
%1:sgpr(s32) = COPY $sgpr1 %1:sgpr(s32) = COPY $sgpr1
%2:sgpr(s32) = G_XOR %0, %1 %2:sgpr(s32) = G_XOR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
name: xor_s64_sgpr_sgpr_sgpr name: xor_s64_sgpr_sgpr_sgpr
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE64-LABEL: name: xor_s64_sgpr_sgpr_sgpr ; WAVE64-LABEL: name: xor_s64_sgpr_sgpr_sgpr
; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 ; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64 = S_XOR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64 = S_XOR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]]
; ;
; WAVE32-LABEL: name: xor_s64_sgpr_sgpr_sgpr ; WAVE32-LABEL: name: xor_s64_sgpr_sgpr_sgpr
; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3 ; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
; WAVE32-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64 = S_XOR_B64 [[COPY]], [[COPY1]], implicit-def $scc ; WAVE32-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64 = S_XOR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]]
%0:sgpr(s64) = COPY $sgpr0_sgpr1 %0:sgpr(s64) = COPY $sgpr0_sgpr1
%1:sgpr(s64) = COPY $sgpr2_sgpr3 %1:sgpr(s64) = COPY $sgpr2_sgpr3
%2:sgpr(s64) = G_XOR %0, %1 %2:sgpr(s64) = G_XOR %0, %1
S_ENDPGM 0, implicit %2 S_ENDPGM 0, implicit %2
... ...
--- ---
▲ Show 20 Lines • Show All 198 Lines • ▼ Show 20 Lines bb.0:
; WAVE64: liveins: $vgpr0, $vgpr1 ; WAVE64: liveins: $vgpr0, $vgpr1
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]]
; ;
; WAVE32-LABEL: name: xor_s1_vcc_copy_to_vcc ; WAVE32-LABEL: name: xor_s1_vcc_copy_to_vcc
; WAVE32: liveins: $vgpr0, $vgpr1 ; WAVE32: liveins: $vgpr0, $vgpr1
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_XOR_B32_]]
%0:vgpr(s32) = COPY $vgpr0 %0:vgpr(s32) = COPY $vgpr0
%1:vgpr(s32) = COPY $vgpr1 %1:vgpr(s32) = COPY $vgpr1
%2:vgpr(s1) = G_TRUNC %0 %2:vgpr(s1) = G_TRUNC %0
%3:vgpr(s1) = G_TRUNC %1 %3:vgpr(s1) = G_TRUNC %1
%4:vcc(s1) = COPY %2 %4:vcc(s1) = COPY %2
%5:vcc(s1) = COPY %3 %5:vcc(s1) = COPY %3
%6:vcc(s1) = G_XOR %4, %5 %6:vcc(s1) = G_XOR %4, %5
Show All 17 Lines bb.0:
; WAVE64: liveins: $vgpr0, $sgpr0 ; WAVE64: liveins: $vgpr0, $sgpr0
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_XOR_B64_]] ; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_XOR_B64_]]
; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]]
; ;
; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave32 ; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
; WAVE32: liveins: $vgpr0, $sgpr0 ; WAVE32: liveins: $vgpr0, $sgpr0
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_XOR_B32_]] ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_XOR_B32_]]
; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]]
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%0:vgpr(s1) = G_TRUNC %1(s32) %0:vgpr(s1) = G_TRUNC %1(s32)
%sgpr0:sgpr(s32) = COPY $sgpr0 %sgpr0:sgpr(s32) = COPY $sgpr0
%2:sgpr(s1) = G_TRUNC %sgpr0 %2:sgpr(s1) = G_TRUNC %sgpr0
%6:sgpr(s32) = G_CONSTANT i32 0 %6:sgpr(s32) = G_CONSTANT i32 0
%7:sgpr(p1) = G_IMPLICIT_DEF %7:sgpr(p1) = G_IMPLICIT_DEF
Show All 19 Lines bb.0:
; WAVE64: liveins: $vgpr0, $sgpr0 ; WAVE64: liveins: $vgpr0, $sgpr0
; WAVE64-NEXT: {{ $}} ; WAVE64-NEXT: {{ $}}
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE64-NEXT: [[S_XOR_B64_:%[0-9]+]]:sreg_64_xexec = S_XOR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]] ; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_XOR_B64_]]
; ;
; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave64 ; WAVE32-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
; WAVE32: liveins: $vgpr0, $sgpr0 ; WAVE32: liveins: $vgpr0, $sgpr0
; WAVE32-NEXT: {{ $}} ; WAVE32-NEXT: {{ $}}
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0 ; WAVE32-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec ; WAVE32-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc ; WAVE32-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec ; WAVE32-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def $scc ; WAVE32-NEXT: [[S_XOR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_XOR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_XOR_B32_]] ; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_XOR_B32_]]
; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]] ; WAVE32-NEXT: S_ENDPGM 0, implicit [[COPY1]]
%1:vgpr(s32) = COPY $vgpr0 %1:vgpr(s32) = COPY $vgpr0
%0:vgpr(s1) = G_TRUNC %1(s32) %0:vgpr(s1) = G_TRUNC %1(s32)
%sgpr0:sgpr(s32) = COPY $sgpr0 %sgpr0:sgpr(s32) = COPY $sgpr0
%2:sgpr(s1) = G_TRUNC %sgpr0 %2:sgpr(s1) = G_TRUNC %sgpr0
%6:sgpr(s32) = G_CONSTANT i32 0 %6:sgpr(s32) = G_CONSTANT i32 0
%7:sgpr(p1) = G_IMPLICIT_DEF %7:sgpr(p1) = G_IMPLICIT_DEF
%9:vcc(s1) = COPY %0(s1) %9:vcc(s1) = COPY %0(s1)
%10:vcc(s1) = COPY %2(s1) %10:vcc(s1) = COPY %2(s1)
%8:vcc(s1) = G_XOR %9, %10 %8:vcc(s1) = G_XOR %9, %10
%3:sreg_64_xexec(s1) = COPY %8(s1) %3:sreg_64_xexec(s1) = COPY %8(s1)
S_ENDPGM 0, implicit %3 S_ENDPGM 0, implicit %3
... ...

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.make.buffer.rsrc.ll

; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 2 ; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 2
; RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -stop-after=instruction-select < %s | FileCheck %s ; RUN: llc -global-isel -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -stop-after=instruction-select < %s | FileCheck %s
define amdgpu_ps ptr addrspace(8) @basic_raw_buffer(ptr inreg %p) { define amdgpu_ps ptr addrspace(8) @basic_raw_buffer(ptr inreg %p) {
; CHECK-LABEL: name: basic_raw_buffer ; CHECK-LABEL: name: basic_raw_buffer
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr0, $sgpr1 ; CHECK-NEXT: liveins: $sgpr0, $sgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5678 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5678
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 1234 ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 1234
; CHECK-NEXT: [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 65535 ; CHECK-NEXT: [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 65535
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_2]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_2]], implicit-def dead $scc
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]] ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_AND_B32_]] ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_AND_B32_]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
Show All 10 Linesdefine amdgpu_ps float @read_raw_buffer(ptr addrspace(1) inreg %p) {
; CHECK-LABEL: name: read_raw_buffer ; CHECK-LABEL: name: read_raw_buffer
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr0, $sgpr1 ; CHECK-NEXT: liveins: $sgpr0, $sgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 65535 ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 65535
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_1]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[S_AND_B32_]], %subreg.sub1, [[S_MOV_B32_]], %subreg.sub2, [[S_MOV_B32_]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[S_AND_B32_]], %subreg.sub1, [[S_MOV_B32_]], %subreg.sub2, [[S_MOV_B32_]], %subreg.sub3
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE]], [[S_MOV_B32_]], 4, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE]], [[S_MOV_B32_]], 4, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%rsrc = call ptr addrspace(8) @llvm.amdgcn.make.buffer.rsrc.p1(ptr addrspace(1) %p, i16 0, i32 0, i32 0) %rsrc = call ptr addrspace(8) @llvm.amdgcn.make.buffer.rsrc.p1(ptr addrspace(1) %p, i16 0, i32 0, i32 0)
%loaded = call float @llvm.amdgcn.raw.ptr.buffer.load(ptr addrspace(8) %rsrc, i32 4, i32 0, i32 0) %loaded = call float @llvm.amdgcn.raw.ptr.buffer.load(ptr addrspace(8) %rsrc, i32 4, i32 0, i32 0)
ret float %loaded ret float %loaded
} }
define amdgpu_ps ptr addrspace(8) @basic_struct_buffer(ptr inreg %p) { define amdgpu_ps ptr addrspace(8) @basic_struct_buffer(ptr inreg %p) {
; CHECK-LABEL: name: basic_struct_buffer ; CHECK-LABEL: name: basic_struct_buffer
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr0, $sgpr1 ; CHECK-NEXT: liveins: $sgpr0, $sgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5678 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5678
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 1234 ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 1234
; CHECK-NEXT: [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 65535 ; CHECK-NEXT: [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 65535
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_2]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_2]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B32_3:%[0-9]+]]:sreg_32 = S_MOV_B32 262144 ; CHECK-NEXT: [[S_MOV_B32_3:%[0-9]+]]:sreg_32 = S_MOV_B32 262144
; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_MOV_B32_3]], implicit-def $scc ; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_MOV_B32_3]], implicit-def dead $scc
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]] ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_OR_B32_]] ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[S_OR_B32_]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
Show All 11 Linesdefine amdgpu_ps ptr addrspace(8) @variable_top_half(ptr inreg %p, i32 inreg %numVals, i32 inreg %flags) {
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3 ; CHECK-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 65535 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 65535
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 262144 ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 262144
; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_MOV_B32_1]], implicit-def $scc ; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]] ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_OR_B32_]] ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_OR_B32_]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY2]] ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY2]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
Show All 12 Linesdefine amdgpu_ps ptr addrspace(8) @general_case(ptr inreg %p, i16 inreg %stride, i32 inreg %numVals, i32 inreg %flags) {
; CHECK-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4 ; CHECK-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 65535 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 65535
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; CHECK-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY2]], [[S_MOV_B32_1]], implicit-def $scc ; CHECK-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY2]], [[S_MOV_B32_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_LSHL_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_LSHL_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[COPY]] ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[COPY]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; CHECK-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_OR_B32_]] ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_OR_B32_]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; CHECK-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY3]] ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY3]]
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
Show All 12 Linesdefine amdgpu_ps float @general_case_load(ptr inreg %p, i16 inreg %stride, i32 inreg %numVals, i32 inreg %flags) {
; CHECK-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4 ; CHECK-NEXT: liveins: $sgpr0, $sgpr1, $sgpr2, $sgpr3, $sgpr4
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 65535 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 65535
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 [[COPY1]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; CHECK-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY2]], [[S_MOV_B32_1]], implicit-def $scc ; CHECK-NEXT: [[S_LSHL_B32_:%[0-9]+]]:sreg_32 = S_LSHL_B32 [[COPY2]], [[S_MOV_B32_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_LSHL_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[S_AND_B32_]], [[S_LSHL_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; CHECK-NEXT: [[S_MOV_B32_2:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[S_OR_B32_]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[S_OR_B32_]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_2]] ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_2]]
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[COPY5]], [[REG_SEQUENCE]], [[S_MOV_B32_2]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[COPY5]], [[REG_SEQUENCE]], [[S_MOV_B32_2]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_IDXEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_IDXEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%rsrc = call ptr addrspace(8) @llvm.amdgcn.make.buffer.rsrc.p0(ptr %p, i16 %stride, i32 %numVals, i32 %flags) %rsrc = call ptr addrspace(8) @llvm.amdgcn.make.buffer.rsrc.p0(ptr %p, i16 %stride, i32 %numVals, i32 %flags)
%value = call float @llvm.amdgcn.struct.ptr.buffer.load(ptr addrspace(8) %rsrc, i32 0, i32 0, i32 0, i32 0) %value = call float @llvm.amdgcn.struct.ptr.buffer.load(ptr addrspace(8) %rsrc, i32 0, i32 0, i32 0, i32 0)
Show All 32 Linesdefine amdgpu_ps float @general_case_load_with_waterfall(ptr %p, i16 %stride, i32 %numVals, i32 %flags) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[COPY7]], [[REG_SEQUENCE1]], [[S_MOV_B32_2]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_IDXEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_IDXEN [[COPY7]], [[REG_SEQUENCE1]], [[S_MOV_B32_2]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 18 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.atomic.add.ll

Show First 20 Lines • Show All 118 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_buffer_atomic_add_i32__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_OFFEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_OFFEN_RTN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_OFFEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_OFFEN_RTN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 39 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_i32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_ATOMIC_ADD_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_ADD_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 60 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.atomic.cmpswap.ll

Show First 20 Lines • Show All 77 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_buffer_atomic_cmpswap_i32__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; CHECK-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN]].sub0 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN]].sub0
▲ Show 20 Lines • Show All 43 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_atomic_cmpswap_i32_noret__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1
; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_OFFEN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_OFFEN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 56 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.atomic.fadd.ll

Show All 12 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_4095__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_4095__sgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_4095__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 16 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 34 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 15 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 13 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY7:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY7:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY8:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY8:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 19 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add4095(float %val, <4 x i32> inreg %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add4095 ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add4095
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; GFX90A-LABEL: name: raw_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<2 x half> %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 16 Linesdefine amdgpu_ps void @raw_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(<2 x half> %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 13 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.load.format.f16.ll

Show All 12 Linesdefine amdgpu_ps half @raw_buffer_load_format_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]] ; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: raw_buffer_load_format_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: raw_buffer_load_format_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps <2 x half> @raw_buffer_load_format_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_XY_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_XY_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>), align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN]] ; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: raw_buffer_load_format_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: raw_buffer_load_format_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 37 Linesdefine amdgpu_ps <4 x half> @raw_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY6]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY6]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY7]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY7]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: raw_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: raw_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 54 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @raw_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.4: ; PACKED-NEXT: bb.4:
; PACKED-NEXT: successors: %bb.5(0x80000000) ; PACKED-NEXT: successors: %bb.5(0x80000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.5: ; PACKED-NEXT: bb.5:
; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]] ; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: raw_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; UNPACKED-LABEL: name: raw_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: successors: %bb.2(0x80000000) ; UNPACKED-NEXT: successors: %bb.2(0x80000000)
; UNPACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; UNPACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps half @raw_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 23 Linesdefine amdgpu_ps <4 x half> @raw_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add_4095(<4 x i32> inreg %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY6]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY6]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY7]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY7]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: raw_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add_4095 ; UNPACKED-LABEL: name: raw_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add_4095
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 38 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.load.format.ll

Show First 20 Lines • Show All 118 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_buffer_load_format_f32__vgpr_rsrc__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 46 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.load.ll

Show First 20 Lines • Show All 67 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_buffer_load_f32__vgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps float @raw_buffer_load_f32__vgpr_rsrc__vgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 378 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @raw_buffer_load_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_USHORT_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_USHORT_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_USHORT_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_USHORT_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps float @raw_buffer_load_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_UBYTE_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_UBYTE_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_UBYTE_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_UBYTE_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8), addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 177 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add16(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%soffset = add i32 %soffset.base, 16 %soffset = add i32 %soffset.base, 16
%val = call float @llvm.amdgcn.raw.buffer.load.f32(<4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 0) %val = call float @llvm.amdgcn.raw.buffer.load.f32(<4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 0)
ret float %val ret float %val
} }
define amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095 ; CHECK-LABEL: name: raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%soffset = add i32 %soffset.base, 4095 %soffset = add i32 %soffset.base, 4095
%val = call float @llvm.amdgcn.raw.buffer.load.f32(<4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 0) %val = call float @llvm.amdgcn.raw.buffer.load.f32(<4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 0)
ret float %val ret float %val
} }
define amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096 ; CHECK-LABEL: name: raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%soffset = add i32 %soffset.base, 4096 %soffset = add i32 %soffset.base, 4096
%val = call float @llvm.amdgcn.raw.buffer.load.f32(<4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 0) %val = call float @llvm.amdgcn.raw.buffer.load.f32(<4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 0)
ret float %val ret float %val
} }
; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop. ; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop.
define amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(<4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(<4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000 ; CHECK-LABEL: name: raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: successors: %bb.2(0x80000000) ; CHECK-NEXT: successors: %bb.2(0x80000000)
; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; CHECK-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.2: ; CHECK-NEXT: bb.2:
; CHECK-NEXT: successors: %bb.3(0x80000000) ; CHECK-NEXT: successors: %bb.3(0x80000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 39 Linesdefine amdgpu_ps float @raw_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add5000(<4 x i32> %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 26 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.store.format.f16.ll

Show All 11 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_f16(<4 x i32> inreg %rsrc, half %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_f16 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 16 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__voffset_4095__sgpr_soffset_f16(<4 x i32> inreg %rsrc, half %val, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFSET_exact [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFSET_exact [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__voffset_4095__sgpr_soffset_f16 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__voffset_4095__sgpr_soffset_f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 20 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16(<4 x i32> inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 24 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(<4 x i32> inreg %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY5]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY5]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 42 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(<4 x i32> %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16 ; PACKED-LABEL: name: raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 15 Linesdefine amdgpu_ps void @raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(<4 x i32> %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 23 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4095(<4 x i32> inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4095 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4095
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4096(<4 x i32> inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4096 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4096
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_16(<4 x i32> inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 16, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 16, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_16 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 22 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4095(<4 x i32> inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4095 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4095
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 25 Linesdefine amdgpu_ps void @raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4096(<4 x i32> inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY5]], [[COPY7]], 0, implicit $exec ; UNPACKED-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY5]], [[COPY7]], 0, implicit $exec
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4096 ; PACKED-LABEL: name: raw_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4096
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 48 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096(<4 x i32> %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096 ; PACKED-LABEL: name: raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 18 Linesdefine amdgpu_ps void @raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096(<4 x i32> %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 16 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.store.format.f32.ll

Show First 20 Lines • Show All 133 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f32(<4 x i32> %rsrc, <4 x float> %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[COPY8]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[COPY8]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 154 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f32_add_4096(<4 x i32> %rsrc, <4 x float> %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE1]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 17 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.buffer.store.ll

Show First 20 Lines • Show All 69 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset(<4 x i32> %rsrc, float %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 78 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, float %val, i32 %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 334 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(<4 x i32> %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORDX2_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORDX2_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 240 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_offset_add_5000(<4 x i32> %rsrc, float %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 37 Linesdefine amdgpu_ps void @raw_buffer_store__vgpr_rsrc__vgpr_val__5000_voffset__sgpr_soffset_offset(<4 x i32> %rsrc, float %val, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 23 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.atomic.add.ll

Show First 20 Lines • Show All 118 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_ptr_buffer_atomic_add_i32__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_OFFEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_OFFEN_RTN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_OFFEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_OFFEN_RTN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 39 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_i32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_ATOMIC_ADD_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_ADD_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 60 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.atomic.cmpswap.ll

Show First 20 Lines • Show All 77 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_ptr_buffer_atomic_cmpswap_i32__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN]].sub0 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_ATOMIC_CMPSWAP_OFFEN_RTN]].sub0
▲ Show 20 Lines • Show All 43 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_cmpswap_i32_noret__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY8]], %subreg.sub0, [[COPY9]], %subreg.sub1
; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_OFFEN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_OFFEN [[REG_SEQUENCE2]], [[COPY10]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 56 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.atomic.fadd.ll

Show All 12 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_4095__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_4095__sgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_4095__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 16 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 34 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 15 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 13 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY7:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY7:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY8:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY8:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_OFFSET [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 19 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add4095(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add4095 ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add4095
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<2 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 16 Linesdefine amdgpu_ps void @raw_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(<2 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX908-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_OFFSET [[COPY]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: raw_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 13 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.load.format.f16.ll

Show All 12 Linesdefine amdgpu_ps half @raw_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]] ; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps <2 x half> @raw_ptr_buffer_load_format_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_XY_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_XY_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN]] ; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_XY_OFFEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 37 Linesdefine amdgpu_ps <4 x half> @raw_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY6]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY6]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY7]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY7]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 54 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @raw_ptr_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.4: ; PACKED-NEXT: bb.4:
; PACKED-NEXT: successors: %bb.5(0x80000000) ; PACKED-NEXT: successors: %bb.5(0x80000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.5: ; PACKED-NEXT: bb.5:
; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]] ; PACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_OFFEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: successors: %bb.2(0x80000000) ; UNPACKED-NEXT: successors: %bb.2(0x80000000)
; UNPACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; UNPACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps half @raw_ptr_buffer_load_format_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 23 Linesdefine amdgpu_ps <4 x half> @raw_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add_4095(ptr addrspace(8) inreg %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub0
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_FORMAT_D16_XYZW_OFFEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY6]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY6]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY7]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY7]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add_4095 ; UNPACKED-LABEL: name: raw_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add_4095
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 38 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.load.format.ll

Show First 20 Lines • Show All 118 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_ptr_buffer_load_format_f32__vgpr_rsrc__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 46 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.load.ll

Show First 20 Lines • Show All 67 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_ptr_buffer_load_f32__vgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps float @raw_ptr_buffer_load_f32__vgpr_rsrc__vgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 378 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @raw_ptr_buffer_load_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_USHORT_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_USHORT_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_USHORT_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_USHORT_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps float @raw_ptr_buffer_load_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_UBYTE_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_UBYTE_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8) from %ir.rsrc, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_UBYTE_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_UBYTE_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable load (s8) from %ir.rsrc, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 176 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add16(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%soffset = add i32 %soffset.base, 16 %soffset = add i32 %soffset.base, 16
%val = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 0) %val = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 0)
ret float %val ret float %val
} }
define amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095 ; CHECK-LABEL: name: raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%soffset = add i32 %soffset.base, 4095 %soffset = add i32 %soffset.base, 4095
%val = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 0) %val = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 0)
ret float %val ret float %val
} }
define amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096 ; CHECK-LABEL: name: raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; CHECK-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; CHECK-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
%soffset = add i32 %soffset.base, 4096 %soffset = add i32 %soffset.base, 4096
%val = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 0) %val = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 0)
ret float %val ret float %val
} }
; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop. ; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop.
define amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000 ; CHECK-LABEL: name: raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: successors: %bb.2(0x80000000) ; CHECK-NEXT: successors: %bb.2(0x80000000)
; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY5]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; CHECK-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; CHECK-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.2: ; CHECK-NEXT: bb.2:
; CHECK-NEXT: successors: %bb.3(0x80000000) ; CHECK-NEXT: successors: %bb.3(0x80000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 39 Linesdefine amdgpu_ps float @raw_ptr_buffer_load_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add5000(ptr addrspace(8) %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 26 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.store.format.f16.ll

Show All 11 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_f16(ptr addrspace(8) inreg %rsrc, half %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_f16 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 16 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__voffset_4095__sgpr_soffset_f16(ptr addrspace(8) inreg %rsrc, half %val, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFSET_exact [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_OFFSET_exact [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 4095, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__voffset_4095__sgpr_soffset_f16 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__voffset_4095__sgpr_soffset_f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16(ptr addrspace(8) inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 24 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(ptr addrspace(8) inreg %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY5]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY5]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 42 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(ptr addrspace(8) %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 15 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(ptr addrspace(8) %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 23 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4095(ptr addrspace(8) inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4095 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4095
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4096(ptr addrspace(8) inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY6]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4096 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_soffset4096
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_16(ptr addrspace(8) inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 16, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 16, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_16 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_16
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 22 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4095(ptr addrspace(8) inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4095 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4095
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 25 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4096(ptr addrspace(8) inreg %rsrc, <2 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY5]], [[COPY7]], 0, implicit $exec ; UNPACKED-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY5]], [[COPY7]], 0, implicit $exec
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4096 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__sgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v2f16_voffset_add_4096
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 48 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096(ptr addrspace(8) %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096 ; PACKED-LABEL: name: raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 18 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16_add_4096(ptr addrspace(8) %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 16 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.store.format.f32.ll

Show First 20 Lines • Show All 133 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f32(ptr addrspace(8) %rsrc, <4 x float> %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[COPY8]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[COPY8]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 154 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store_format__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f32_add_4096(ptr addrspace(8) %rsrc, <4 x float> %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_FORMAT_XYZW_OFFEN_exact [[REG_SEQUENCE]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE2]], [[COPY9]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 17 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.buffer.store.ll

Show First 20 Lines • Show All 69 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset(ptr addrspace(8) %rsrc, float %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 78 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, float %val, i32 %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 334 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_v4f16(ptr addrspace(8) %rsrc, <4 x half> %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORDX2_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORDX2_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE2]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 240 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_ptr_buffer_store__vgpr_rsrc__vgpr_val__vgpr_voffset__sgpr_soffset_offset_add_5000(ptr addrspace(8) %rsrc, float %val, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 37 Linesdefine amdgpu_ps void @raw_ptr_buffer_store__vgpr_rsrc__vgpr_val__5000_voffset__sgpr_soffset_offset(ptr addrspace(8) %rsrc, float %val, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORD_OFFEN_exact [[COPY4]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY5]], 904, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 23 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.tbuffer.load.f16.ll

Show All 11 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 28 Linesdefine amdgpu_ps <2 x half> @raw_tbuffer_load_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY7]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY7]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY10]], [[V_AND_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY10]], [[V_AND_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 44 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x half> @raw_tbuffer_load_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY9]], [[COPY14]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY9]], [[COPY14]], implicit $exec
; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY15]], [[V_AND_B32_e64_3]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY15]], [[V_AND_B32_e64_3]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]] ; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; PACKED-LABEL: name: raw_tbuffer_load_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 35 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 20 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 1, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 1, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 2, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 2, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 3, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 3, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 4, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 4, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 16 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.tbuffer.load.ll

Show First 20 Lines • Show All 117 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_tbuffer_load_f32__vgpr_rsrc__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 95 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.tbuffer.store.f16.ll

Show All 11 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(half %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 21 Linesdefine amdgpu_ps void @raw_tbuffer_store_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<2 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 30 Linesdefine amdgpu_ps void @raw_tbuffer_store_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY1]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1, [[COPY4]], %subreg.sub2, [[COPY5]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1, [[COPY4]], %subreg.sub2, [[COPY5]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY7]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE]], [[COPY6]], [[REG_SEQUENCE1]], [[COPY7]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset(half %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset(half %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset(half %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset(half %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(half %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 14 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(half %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 19 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc(half %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 1, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 1, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(half %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 2, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 2, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc(half %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 3, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 3, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc(half %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 4, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 4, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 14 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.tbuffer.store.i8.ll

Show All 11 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__sgpr_rsrc__vgpr_voffset__sgpr_soffset(i8 %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 33 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset(i8 %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset(i8 %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset(i8 %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset(i8 %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i8 %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 14 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i8 %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8) into %ir.rsrc, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 13 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.ptr.tbuffer.store.ll

Show First 20 Lines • Show All 137 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__vgpr_rsrc__vgpr_voffset__sgpr_soffset(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 1, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 1, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__vgpr_rsrc__vgpr_voffset__vgpr_soffset(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float %val, ptr addrspace(8) %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 256 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add16(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: S_ENDPGM 0 ; CHECK-NEXT: S_ENDPGM 0
%soffset = add i32 %soffset.base, 16 %soffset = add i32 %soffset.base, 16
call void @llvm.amdgcn.raw.ptr.tbuffer.store.f32(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0) call void @llvm.amdgcn.raw.ptr.tbuffer.store.f32(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0)
ret void ret void
} }
define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095 ; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: S_ENDPGM 0 ; CHECK-NEXT: S_ENDPGM 0
%soffset = add i32 %soffset.base, 4095 %soffset = add i32 %soffset.base, 4095
call void @llvm.amdgcn.raw.ptr.tbuffer.store.f32(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0) call void @llvm.amdgcn.raw.ptr.tbuffer.store.f32(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0)
ret void ret void
} }
define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(float %val, ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096 ; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: S_ENDPGM 0 ; CHECK-NEXT: S_ENDPGM 0
%soffset = add i32 %soffset.base, 4096 %soffset = add i32 %soffset.base, 4096
call void @llvm.amdgcn.raw.ptr.tbuffer.store.f32(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0) call void @llvm.amdgcn.raw.ptr.tbuffer.store.f32(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0)
ret void ret void
} }
; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop. ; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop.
define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(float %val, ptr addrspace(8) %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000 ; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: successors: %bb.2(0x80000000) ; CHECK-NEXT: successors: %bb.2(0x80000000)
; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_MOV_B32 $exec_lo ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_MOV_B32 $exec_lo
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.2: ; CHECK-NEXT: bb.2:
; CHECK-NEXT: successors: %bb.3(0x80000000) ; CHECK-NEXT: successors: %bb.3(0x80000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 39 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add5000(float %val, ptr addrspace(8) %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[V_ADD_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[V_ADD_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 94, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 17 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.tbuffer.load.f16.ll

Show All 11 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 28 Linesdefine amdgpu_ps <2 x half> @raw_tbuffer_load_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY7]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY7]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY10]], [[V_AND_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY10]], [[V_AND_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 44 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x half> @raw_tbuffer_load_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY9]], [[COPY14]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY9]], [[COPY14]], implicit $exec
; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY15]], [[V_AND_B32_e64_3]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY15]], [[V_AND_B32_e64_3]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]] ; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; PACKED-LABEL: name: raw_tbuffer_load_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 35 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 20 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 1, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 1, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 2, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 2, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 3, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 3, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps half @raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 4, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 4, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_gfx80_OFFEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc ; PACKED-LABEL: name: raw_tbuffer_load_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 16 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.tbuffer.load.ll

Show First 20 Lines • Show All 117 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @raw_tbuffer_load_f32__vgpr_rsrc__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY5]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_X_OFFEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_X_OFFEN [[COPY6]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 95 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.tbuffer.store.f16.ll

Show All 11 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(half %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 21 Linesdefine amdgpu_ps void @raw_tbuffer_store_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<2 x half> %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY7]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XY_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_v2f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 30 Linesdefine amdgpu_ps void @raw_tbuffer_store_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<4 x half> %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY1]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE]], [[COPY7]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_XYZW_gfx80_OFFEN_exact [[REG_SEQUENCE1]], [[COPY6]], [[REG_SEQUENCE]], [[COPY7]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_v4f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset(half %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__sgpr_soffset(half %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset(half %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__vgpr_voffset__vgpr_soffset(half %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(half %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 14 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__vgpr_rsrc__sgpr_voffset__vgpr_soffset(half %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 19 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc(half %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 1, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 1, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(half %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 2, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 2, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc(half %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 3, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 3, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc_glc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Linesdefine amdgpu_ps void @raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc(half %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 4, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_D16_X_gfx80_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 4, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc ; PACKED-LABEL: name: raw_tbuffer_store_f16__sgpr_rsrc__vgpr_voffset__sgpr_soffset_dlc
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 14 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.tbuffer.store.i8.ll

Show All 11 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__sgpr_rsrc__vgpr_voffset__sgpr_soffset(i8 %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 33 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset(i8 %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__sgpr_soffset(i8 %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset(i8 %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6 ; PACKED-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__vgpr_voffset__vgpr_soffset(i8 %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i8 %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; UNPACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; PACKED-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 14 Linesdefine amdgpu_ps void @raw_tbuffer_store_i8__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i8 %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8) ; PACKED-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s8), addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 13 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.raw.tbuffer.store.ll

Show First 20 Lines • Show All 137 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__vgpr_rsrc__vgpr_voffset__sgpr_soffset(float %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 1, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[COPY6]], 0, 94, 1, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__vgpr_rsrc__vgpr_voffset__vgpr_soffset(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float %val, <4 x i32> %rsrc, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[S_AND_B32_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY7]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
▲ Show 20 Lines • Show All 257 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add16(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: S_ENDPGM 0 ; CHECK-NEXT: S_ENDPGM 0
%soffset = add i32 %soffset.base, 16 %soffset = add i32 %soffset.base, 16
call void @llvm.amdgcn.raw.tbuffer.store.f32(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0) call void @llvm.amdgcn.raw.tbuffer.store.f32(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0)
ret void ret void
} }
define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095 ; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4095
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4095
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: S_ENDPGM 0 ; CHECK-NEXT: S_ENDPGM 0
%soffset = add i32 %soffset.base, 4095 %soffset = add i32 %soffset.base, 4095
call void @llvm.amdgcn.raw.tbuffer.store.f32(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0) call void @llvm.amdgcn.raw.tbuffer.store.f32(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0)
ret void ret void
} }
define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096(float %val, <4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096 ; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add4096
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; CHECK-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: S_ENDPGM 0 ; CHECK-NEXT: S_ENDPGM 0
%soffset = add i32 %soffset.base, 4096 %soffset = add i32 %soffset.base, 4096
call void @llvm.amdgcn.raw.tbuffer.store.f32(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0) call void @llvm.amdgcn.raw.tbuffer.store.f32(float %val, <4 x i32> %rsrc, i32 %voffset, i32 %soffset, i32 94, i32 0)
ret void ret void
} }
; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop. ; An add of the offset is necessary, with a waterfall loop. Make sure the add is done outside of the waterfall loop.
define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(float %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset.base) { define amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000(float %val, <4 x i32> %rsrc, i32 %voffset, i32 inreg %soffset.base) {
; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000 ; CHECK-LABEL: name: raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_soffset_add5000
; CHECK: bb.1 (%ir-block.0): ; CHECK: bb.1 (%ir-block.0):
; CHECK-NEXT: successors: %bb.2(0x80000000) ; CHECK-NEXT: successors: %bb.2(0x80000000)
; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5 ; CHECK-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4 ; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr4
; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5 ; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr5
; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; CHECK-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr2
; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000 ; CHECK-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 5000
; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def $scc ; CHECK-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY6]], [[S_MOV_B32_]], implicit-def dead $scc
; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_MOV_B32 $exec_lo ; CHECK-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0_xexec = S_MOV_B32 $exec_lo
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.2: ; CHECK-NEXT: bb.2:
; CHECK-NEXT: successors: %bb.3(0x80000000) ; CHECK-NEXT: successors: %bb.3(0x80000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[COPY5]], [[REG_SEQUENCE1]], [[S_ADD_I32_]], 0, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 39 Linesdefine amdgpu_ps void @raw_tbuffer_store_f32__sgpr_rsrc__vgpr_voffset__sgpr_soffset_voffset_add5000(float %val, <4 x i32> %rsrc, i32 %voffset.base, i32 inreg %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY8:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY10]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_B32 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_AND_SAVEEXEC_B32 killed [[S_AND_B32_]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[V_ADD_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: TBUFFER_STORE_FORMAT_X_OFFEN_exact [[COPY]], [[V_ADD_U32_e64_]], [[REG_SEQUENCE1]], [[COPY6]], 904, 94, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc ; CHECK-NEXT: $exec_lo = S_XOR_B32_term $exec_lo, [[S_AND_SAVEEXEC_B32_]], implicit-def $scc
; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; CHECK-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 17 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.s.buffer.load.ll

  • This file is larger than 256 KB, so syntax highlighting is disabled by default.
Show All 16 Linesdefine amdgpu_ps i32 @s_buffer_load_i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32 ; GFX7-LABEL: name: s_buffer_load_i32
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32 ; GFX8-LABEL: name: s_buffer_load_i32
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_glc(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_glc ; GFX7-LABEL: name: s_buffer_load_i32_glc
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[COPY4]], 1 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_glc ; GFX8-LABEL: name: s_buffer_load_i32_glc
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 24 Linesdefine amdgpu_ps <2 x i32> @s_buffer_load_v2i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]] ; GFX6-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
; GFX6-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; GFX6-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
;
; GFX7-LABEL: name: s_buffer_load_v2i32 ; GFX7-LABEL: name: s_buffer_load_v2i32
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s64), align 4) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORDX2_SGPR:%[0-9]+]]:sreg_64_xexec = S_BUFFER_LOAD_DWORDX2_SGPR [[REG_SEQUENCE]], [[COPY4]], 0 :: (dereferenceable invariant load (s64), align 4)
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub0 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub0
; GFX7-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY [[S_BUFFER_LOAD_DWORDX2_SGPR]].sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]] ; GFX7-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
; GFX7-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; GFX7-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1
;
; GFX8-LABEL: name: s_buffer_load_v2i32 ; GFX8-LABEL: name: s_buffer_load_v2i32
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 34 Linesdefine amdgpu_ps <3 x i32> @s_buffer_load_v3i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]] ; GFX6-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec
; GFX6-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; GFX6-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; GFX6-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[COPY7]] ; GFX6-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[COPY7]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY11]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY11]], implicit $exec
; GFX6-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]] ; GFX6-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2
;
; GFX7-LABEL: name: s_buffer_load_v3i32 ; GFX7-LABEL: name: s_buffer_load_v3i32
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 9 Linesdefine amdgpu_ps <3 x i32> @s_buffer_load_v3i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]] ; GFX7-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[COPY6]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec
; GFX7-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]] ; GFX7-NEXT: $sgpr1 = COPY [[V_READFIRSTLANE_B32_1]]
; GFX7-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[COPY7]] ; GFX7-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[COPY7]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY11]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY11]], implicit $exec
; GFX7-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]] ; GFX7-NEXT: $sgpr2 = COPY [[V_READFIRSTLANE_B32_2]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2
;
; GFX8-LABEL: name: s_buffer_load_v3i32 ; GFX8-LABEL: name: s_buffer_load_v3i32
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 58 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: $sgpr5 = COPY [[V_READFIRSTLANE_B32_5]] ; GFX6-NEXT: $sgpr5 = COPY [[V_READFIRSTLANE_B32_5]]
; GFX6-NEXT: [[COPY19:%[0-9]+]]:vgpr_32 = COPY [[COPY11]] ; GFX6-NEXT: [[COPY19:%[0-9]+]]:vgpr_32 = COPY [[COPY11]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_6:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY19]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_6:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY19]], implicit $exec
; GFX6-NEXT: $sgpr6 = COPY [[V_READFIRSTLANE_B32_6]] ; GFX6-NEXT: $sgpr6 = COPY [[V_READFIRSTLANE_B32_6]]
; GFX6-NEXT: [[COPY20:%[0-9]+]]:vgpr_32 = COPY [[COPY12]] ; GFX6-NEXT: [[COPY20:%[0-9]+]]:vgpr_32 = COPY [[COPY12]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_7:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY20]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_7:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY20]], implicit $exec
; GFX6-NEXT: $sgpr7 = COPY [[V_READFIRSTLANE_B32_7]] ; GFX6-NEXT: $sgpr7 = COPY [[V_READFIRSTLANE_B32_7]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8i32 ; GFX7-LABEL: name: s_buffer_load_v8i32
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 28 Linesdefine amdgpu_ps <8 x i32> @s_buffer_load_v8i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX7-NEXT: $sgpr5 = COPY [[V_READFIRSTLANE_B32_5]] ; GFX7-NEXT: $sgpr5 = COPY [[V_READFIRSTLANE_B32_5]]
; GFX7-NEXT: [[COPY19:%[0-9]+]]:vgpr_32 = COPY [[COPY11]] ; GFX7-NEXT: [[COPY19:%[0-9]+]]:vgpr_32 = COPY [[COPY11]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_6:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY19]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_6:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY19]], implicit $exec
; GFX7-NEXT: $sgpr6 = COPY [[V_READFIRSTLANE_B32_6]] ; GFX7-NEXT: $sgpr6 = COPY [[V_READFIRSTLANE_B32_6]]
; GFX7-NEXT: [[COPY20:%[0-9]+]]:vgpr_32 = COPY [[COPY12]] ; GFX7-NEXT: [[COPY20:%[0-9]+]]:vgpr_32 = COPY [[COPY12]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_7:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY20]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_7:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY20]], implicit $exec
; GFX7-NEXT: $sgpr7 = COPY [[V_READFIRSTLANE_B32_7]] ; GFX7-NEXT: $sgpr7 = COPY [[V_READFIRSTLANE_B32_7]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8i32 ; GFX8-LABEL: name: s_buffer_load_v8i32
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 109 Lines • ▼ Show 20 Linesdefine amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: $sgpr13 = COPY [[V_READFIRSTLANE_B32_13]] ; GFX6-NEXT: $sgpr13 = COPY [[V_READFIRSTLANE_B32_13]]
; GFX6-NEXT: [[COPY35:%[0-9]+]]:vgpr_32 = COPY [[COPY19]] ; GFX6-NEXT: [[COPY35:%[0-9]+]]:vgpr_32 = COPY [[COPY19]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_14:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY35]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_14:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY35]], implicit $exec
; GFX6-NEXT: $sgpr14 = COPY [[V_READFIRSTLANE_B32_14]] ; GFX6-NEXT: $sgpr14 = COPY [[V_READFIRSTLANE_B32_14]]
; GFX6-NEXT: [[COPY36:%[0-9]+]]:vgpr_32 = COPY [[COPY20]] ; GFX6-NEXT: [[COPY36:%[0-9]+]]:vgpr_32 = COPY [[COPY20]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_15:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY36]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_15:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY36]], implicit $exec
; GFX6-NEXT: $sgpr15 = COPY [[V_READFIRSTLANE_B32_15]] ; GFX6-NEXT: $sgpr15 = COPY [[V_READFIRSTLANE_B32_15]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15
;
; GFX7-LABEL: name: s_buffer_load_v16i32 ; GFX7-LABEL: name: s_buffer_load_v16i32
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 60 Lines • ▼ Show 20 Linesdefine amdgpu_ps <16 x i32> @s_buffer_load_v16i32(<4 x i32> inreg %rsrc, i32 inreg %soffset) {
; GFX7-NEXT: $sgpr13 = COPY [[V_READFIRSTLANE_B32_13]] ; GFX7-NEXT: $sgpr13 = COPY [[V_READFIRSTLANE_B32_13]]
; GFX7-NEXT: [[COPY35:%[0-9]+]]:vgpr_32 = COPY [[COPY19]] ; GFX7-NEXT: [[COPY35:%[0-9]+]]:vgpr_32 = COPY [[COPY19]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_14:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY35]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_14:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY35]], implicit $exec
; GFX7-NEXT: $sgpr14 = COPY [[V_READFIRSTLANE_B32_14]] ; GFX7-NEXT: $sgpr14 = COPY [[V_READFIRSTLANE_B32_14]]
; GFX7-NEXT: [[COPY36:%[0-9]+]]:vgpr_32 = COPY [[COPY20]] ; GFX7-NEXT: [[COPY36:%[0-9]+]]:vgpr_32 = COPY [[COPY20]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_15:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY36]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_15:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY36]], implicit $exec
; GFX7-NEXT: $sgpr15 = COPY [[V_READFIRSTLANE_B32_15]] ; GFX7-NEXT: $sgpr15 = COPY [[V_READFIRSTLANE_B32_15]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0, implicit $sgpr1, implicit $sgpr2, implicit $sgpr3, implicit $sgpr4, implicit $sgpr5, implicit $sgpr6, implicit $sgpr7, implicit $sgpr8, implicit $sgpr9, implicit $sgpr10, implicit $sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15
;
; GFX8-LABEL: name: s_buffer_load_v16i32 ; GFX8-LABEL: name: s_buffer_load_v16i32
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 80 Lines • ▼ Show 20 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_1(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_1 ; GFX7-LABEL: name: s_buffer_load_i32_offset_1
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_1 ; GFX8-LABEL: name: s_buffer_load_i32_offset_1
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_glc_4(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_glc_4 ; GFX7-LABEL: name: s_buffer_load_i32_offset_glc_4
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 1, 1 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_glc_4 ; GFX8-LABEL: name: s_buffer_load_i32_offset_glc_4
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_255(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 255 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 255
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_255 ; GFX7-LABEL: name: s_buffer_load_i32_offset_255
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 255 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 255
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_255 ; GFX8-LABEL: name: s_buffer_load_i32_offset_255
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_256(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_256 ; GFX7-LABEL: name: s_buffer_load_i32_offset_256
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 64, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_256 ; GFX8-LABEL: name: s_buffer_load_i32_offset_256
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 17 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_1020(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_1020 ; GFX7-LABEL: name: s_buffer_load_i32_offset_1020
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM [[REG_SEQUENCE]], 255, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_1020 ; GFX8-LABEL: name: s_buffer_load_i32_offset_1020
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_1023(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1023 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1023
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_1023 ; GFX7-LABEL: name: s_buffer_load_i32_offset_1023
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1023 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1023
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_1023 ; GFX8-LABEL: name: s_buffer_load_i32_offset_1023
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_1024(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1024
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_1024 ; GFX7-LABEL: name: s_buffer_load_i32_offset_1024
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 256, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 256, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_1024 ; GFX8-LABEL: name: s_buffer_load_i32_offset_1024
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_1025(<4 x i32> inreg %rsrc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1025 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1025
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_1025 ; GFX7-LABEL: name: s_buffer_load_i32_offset_1025
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1025 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1025
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_1025 ; GFX8-LABEL: name: s_buffer_load_i32_offset_1025
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_neg1(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_neg1 ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg1
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_neg1 ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg1
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_neg4(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -4 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -4
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_neg4 ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg4
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741823, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741823, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_neg4 ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg4
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_neg8(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -8 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -8
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_neg8 ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg8
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741822, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073741822, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_neg8 ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg8
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_bit31(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -2147483648 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -2147483648
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_bit31 ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit31
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 536870912, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 536870912, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_bit31 ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit31
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_glc_bit30(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1073741824 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1073741824
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 1 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 1 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_glc_bit30 ; GFX7-LABEL: name: s_buffer_load_i32_offset_glc_bit30
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 268435456, 1 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 268435456, 1 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_glc_bit30 ; GFX8-LABEL: name: s_buffer_load_i32_offset_glc_bit30
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_bit29(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 536870912 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 536870912
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_bit29 ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit29
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 134217728, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 134217728, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_bit29 ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit29
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_bit21(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 2097152 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 2097152
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_bit21 ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit21
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 524288, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 524288, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_bit21 ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit21
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_bit20(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1048576 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 1048576
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_bit20 ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit20
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 262144, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 262144, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_bit20 ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit20
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_neg_bit20(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1048576 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -1048576
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_neg_bit20 ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg_bit20
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073479680, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073479680, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_neg_bit20 ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg_bit20
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_bit19(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 524288 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 524288
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_bit19 ; GFX7-LABEL: name: s_buffer_load_i32_offset_bit19
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 131072, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 131072, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_bit19 ; GFX8-LABEL: name: s_buffer_load_i32_offset_bit19
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps i32 @s_buffer_load_i32_offset_neg_bit19(<4 x i32> inreg %desc) {
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -524288 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 -524288
; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[S_BUFFER_LOAD_DWORD_SGPR:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_SGPR [[REG_SEQUENCE]], [[S_MOV_B32_]], 0 :: (dereferenceable invariant load (s32))
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]] ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_SGPR]]
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX6-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX7-LABEL: name: s_buffer_load_i32_offset_neg_bit19 ; GFX7-LABEL: name: s_buffer_load_i32_offset_neg_bit19
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073610752, 0 :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[S_BUFFER_LOAD_DWORD_IMM_ci:%[0-9]+]]:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM_ci [[REG_SEQUENCE]], 1073610752, 0 :: (dereferenceable invariant load (s32))
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]] ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[S_BUFFER_LOAD_DWORD_IMM_ci]]
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]] ; GFX7-NEXT: $sgpr0 = COPY [[V_READFIRSTLANE_B32_]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $sgpr0
;
; GFX8-LABEL: name: s_buffer_load_i32_offset_neg_bit19 ; GFX8-LABEL: name: s_buffer_load_i32_offset_neg_bit19
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 20 Linesdefine amdgpu_ps <2 x float> @s_buffer_load_v2f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4)
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1
; GFX6-NEXT: $vgpr0 = COPY [[COPY5]] ; GFX6-NEXT: $vgpr0 = COPY [[COPY5]]
; GFX6-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; GFX7-LABEL: name: s_buffer_load_v2f32_vgpr_offset ; GFX7-LABEL: name: s_buffer_load_v2f32_vgpr_offset
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX2_OFFEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_DWORDX2_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s64), align 4)
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub0
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX2_OFFEN]].sub1
; GFX7-NEXT: $vgpr0 = COPY [[COPY5]] ; GFX7-NEXT: $vgpr0 = COPY [[COPY5]]
; GFX7-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; GFX8-LABEL: name: s_buffer_load_v2f32_vgpr_offset ; GFX8-LABEL: name: s_buffer_load_v2f32_vgpr_offset
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 26 Linesdefine amdgpu_ps <3 x float> @s_buffer_load_v3f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
; GFX6-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
; GFX6-NEXT: $vgpr0 = COPY [[COPY5]] ; GFX6-NEXT: $vgpr0 = COPY [[COPY5]]
; GFX6-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
;
; GFX7-LABEL: name: s_buffer_load_v3f32_vgpr_offset ; GFX7-LABEL: name: s_buffer_load_v3f32_vgpr_offset
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
; GFX7-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
; GFX7-NEXT: $vgpr0 = COPY [[COPY5]] ; GFX7-NEXT: $vgpr0 = COPY [[COPY5]]
; GFX7-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
;
; GFX8-LABEL: name: s_buffer_load_v3f32_vgpr_offset ; GFX8-LABEL: name: s_buffer_load_v3f32_vgpr_offset
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 30 Linesdefine amdgpu_ps <4 x float> @s_buffer_load_v4f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
; GFX6-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
; GFX6-NEXT: $vgpr0 = COPY [[COPY5]] ; GFX6-NEXT: $vgpr0 = COPY [[COPY5]]
; GFX6-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
;
; GFX7-LABEL: name: s_buffer_load_v4f32_vgpr_offset ; GFX7-LABEL: name: s_buffer_load_v4f32_vgpr_offset
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub0
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub2
; GFX7-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[BUFFER_LOAD_DWORDX4_OFFEN]].sub3
; GFX7-NEXT: $vgpr0 = COPY [[COPY5]] ; GFX7-NEXT: $vgpr0 = COPY [[COPY5]]
; GFX7-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
;
; GFX8-LABEL: name: s_buffer_load_v4f32_vgpr_offset ; GFX8-LABEL: name: s_buffer_load_v4f32_vgpr_offset
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 41 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY9]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY9]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY12]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 15 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY9]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY9]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY12]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 69 Lines • ▼ Show 20 Linesdefine amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX6-NEXT: $vgpr9 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr9 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr10 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr10 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr11 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr11 = COPY [[COPY16]]
; GFX6-NEXT: $vgpr12 = COPY [[COPY17]] ; GFX6-NEXT: $vgpr12 = COPY [[COPY17]]
; GFX6-NEXT: $vgpr13 = COPY [[COPY18]] ; GFX6-NEXT: $vgpr13 = COPY [[COPY18]]
; GFX6-NEXT: $vgpr14 = COPY [[COPY19]] ; GFX6-NEXT: $vgpr14 = COPY [[COPY19]]
; GFX6-NEXT: $vgpr15 = COPY [[COPY20]] ; GFX6-NEXT: $vgpr15 = COPY [[COPY20]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
;
; GFX7-LABEL: name: s_buffer_load_v16f32_vgpr_offset ; GFX7-LABEL: name: s_buffer_load_v16f32_vgpr_offset
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 33 Linesdefine amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset(<4 x i32> inreg %rsrc, i32 %soffset) {
; GFX7-NEXT: $vgpr9 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr9 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr10 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr10 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr11 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr11 = COPY [[COPY16]]
; GFX7-NEXT: $vgpr12 = COPY [[COPY17]] ; GFX7-NEXT: $vgpr12 = COPY [[COPY17]]
; GFX7-NEXT: $vgpr13 = COPY [[COPY18]] ; GFX7-NEXT: $vgpr13 = COPY [[COPY18]]
; GFX7-NEXT: $vgpr14 = COPY [[COPY19]] ; GFX7-NEXT: $vgpr14 = COPY [[COPY19]]
; GFX7-NEXT: $vgpr15 = COPY [[COPY20]] ; GFX7-NEXT: $vgpr15 = COPY [[COPY20]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
;
; GFX8-LABEL: name: s_buffer_load_v16f32_vgpr_offset ; GFX8-LABEL: name: s_buffer_load_v16f32_vgpr_offset
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 52 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4092(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4092
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4095(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4095
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_offset_add_4096(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_offset_add_4096
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 36 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY9]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY9]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY12]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4064 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4064
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 15 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4064(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY9]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY9]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY12]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4064 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4064
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 53 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY9]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY9]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY12]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4068 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4068
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 15 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_add_4068(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY6]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY6]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY7]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY7]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY8]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY8]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY9]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY9]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY12]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4068 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_add_4068
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 70 Lines • ▼ Show 20 Linesdefine amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: $vgpr9 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr9 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr10 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr10 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr11 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr11 = COPY [[COPY16]]
; GFX6-NEXT: $vgpr12 = COPY [[COPY17]] ; GFX6-NEXT: $vgpr12 = COPY [[COPY17]]
; GFX6-NEXT: $vgpr13 = COPY [[COPY18]] ; GFX6-NEXT: $vgpr13 = COPY [[COPY18]]
; GFX6-NEXT: $vgpr14 = COPY [[COPY19]] ; GFX6-NEXT: $vgpr14 = COPY [[COPY19]]
; GFX6-NEXT: $vgpr15 = COPY [[COPY20]] ; GFX6-NEXT: $vgpr15 = COPY [[COPY20]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
;
; GFX7-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4032 ; GFX7-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4032
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 33 Linesdefine amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4032(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX7-NEXT: $vgpr9 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr9 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr10 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr10 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr11 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr11 = COPY [[COPY16]]
; GFX7-NEXT: $vgpr12 = COPY [[COPY17]] ; GFX7-NEXT: $vgpr12 = COPY [[COPY17]]
; GFX7-NEXT: $vgpr13 = COPY [[COPY18]] ; GFX7-NEXT: $vgpr13 = COPY [[COPY18]]
; GFX7-NEXT: $vgpr14 = COPY [[COPY19]] ; GFX7-NEXT: $vgpr14 = COPY [[COPY19]]
; GFX7-NEXT: $vgpr15 = COPY [[COPY20]] ; GFX7-NEXT: $vgpr15 = COPY [[COPY20]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
;
; GFX8-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4032 ; GFX8-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4032
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 88 Lines • ▼ Show 20 Linesdefine amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX6-NEXT: $vgpr9 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr9 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr10 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr10 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr11 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr11 = COPY [[COPY16]]
; GFX6-NEXT: $vgpr12 = COPY [[COPY17]] ; GFX6-NEXT: $vgpr12 = COPY [[COPY17]]
; GFX6-NEXT: $vgpr13 = COPY [[COPY18]] ; GFX6-NEXT: $vgpr13 = COPY [[COPY18]]
; GFX6-NEXT: $vgpr14 = COPY [[COPY19]] ; GFX6-NEXT: $vgpr14 = COPY [[COPY19]]
; GFX6-NEXT: $vgpr15 = COPY [[COPY20]] ; GFX6-NEXT: $vgpr15 = COPY [[COPY20]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
;
; GFX7-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4036 ; GFX7-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4036
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 33 Linesdefine amdgpu_ps <16 x float> @s_buffer_load_v16f32_vgpr_offset_add_4036(<4 x i32> inreg %rsrc, i32 %soffset.base) {
; GFX7-NEXT: $vgpr9 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr9 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr10 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr10 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr11 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr11 = COPY [[COPY16]]
; GFX7-NEXT: $vgpr12 = COPY [[COPY17]] ; GFX7-NEXT: $vgpr12 = COPY [[COPY17]]
; GFX7-NEXT: $vgpr13 = COPY [[COPY18]] ; GFX7-NEXT: $vgpr13 = COPY [[COPY18]]
; GFX7-NEXT: $vgpr14 = COPY [[COPY19]] ; GFX7-NEXT: $vgpr14 = COPY [[COPY19]]
; GFX7-NEXT: $vgpr15 = COPY [[COPY20]] ; GFX7-NEXT: $vgpr15 = COPY [[COPY20]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
;
; GFX8-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4036 ; GFX8-LABEL: name: s_buffer_load_v16f32_vgpr_offset_add_4036
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 69 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.4: ; GFX6-NEXT: bb.4:
; GFX6-NEXT: successors: %bb.5(0x80000000) ; GFX6-NEXT: successors: %bb.5(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.5: ; GFX6-NEXT: bb.5:
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.4: ; GFX7-NEXT: bb.4:
; GFX7-NEXT: successors: %bb.5(0x80000000) ; GFX7-NEXT: successors: %bb.5(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.5: ; GFX7-NEXT: bb.5:
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc(<4 x i32> %rsrc, i32 inreg %soffset) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 34 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.4: ; GFX6-NEXT: bb.4:
; GFX6-NEXT: successors: %bb.5(0x80000000) ; GFX6-NEXT: successors: %bb.5(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.5: ; GFX6-NEXT: bb.5:
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4092 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4092
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.4: ; GFX7-NEXT: bb.4:
; GFX7-NEXT: successors: %bb.5(0x80000000) ; GFX7-NEXT: successors: %bb.5(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.5: ; GFX7-NEXT: bb.5:
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4092 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4092
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4092(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4092, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 20 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_soffset_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX6-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX6-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX6-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.2: ; GFX6-NEXT: bb.2:
; GFX6-NEXT: successors: %bb.3(0x80000000) ; GFX6-NEXT: successors: %bb.3(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.4: ; GFX6-NEXT: bb.4:
; GFX6-NEXT: successors: %bb.5(0x80000000) ; GFX6-NEXT: successors: %bb.5(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.5: ; GFX6-NEXT: bb.5:
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4096 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4096
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX7-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX7-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX7-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX7-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX7-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.2: ; GFX7-NEXT: bb.2:
; GFX7-NEXT: successors: %bb.3(0x80000000) ; GFX7-NEXT: successors: %bb.3(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.4: ; GFX7-NEXT: bb.4:
; GFX7-NEXT: successors: %bb.5(0x80000000) ; GFX7-NEXT: successors: %bb.5(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.5: ; GFX7-NEXT: bb.5:
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4096 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_soffset_add_4096
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX8-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX8-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX8-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX8-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX8-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX8-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX8-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.2: ; GFX8-NEXT: bb.2:
; GFX8-NEXT: successors: %bb.3(0x80000000) ; GFX8-NEXT: successors: %bb.3(0x80000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.4: ; GFX6-NEXT: bb.4:
; GFX6-NEXT: successors: %bb.5(0x80000000) ; GFX6-NEXT: successors: %bb.5(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.5: ; GFX6-NEXT: bb.5:
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4095 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4095
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.4: ; GFX7-NEXT: bb.4:
; GFX7-NEXT: successors: %bb.5(0x80000000) ; GFX7-NEXT: successors: %bb.5(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.5: ; GFX7-NEXT: bb.5:
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFSET]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4095 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4095
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4095(<4 x i32> %rsrc) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1) ; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4095, align 1)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 36 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX6-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.4: ; GFX6-NEXT: bb.4:
; GFX6-NEXT: successors: %bb.5(0x80000000) ; GFX6-NEXT: successors: %bb.5(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX6-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.5: ; GFX6-NEXT: bb.5:
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4096 ; GFX7-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4096
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 13 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX7-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.4: ; GFX7-NEXT: bb.4:
; GFX7-NEXT: successors: %bb.5(0x80000000) ; GFX7-NEXT: successors: %bb.5(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX7-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.5: ; GFX7-NEXT: bb.5:
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4096 ; GFX8-LABEL: name: s_buffer_load_f32_vgpr_rsrc_offset_4096
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps float @s_buffer_load_f32_vgpr_rsrc_offset_4096(<4 x i32> %rsrc) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4096) ; GFX8-NEXT: [[BUFFER_LOAD_DWORD_OFFSET:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4095, 0, 0, implicit $exec :: (dereferenceable invariant load (s32) from unknown-address + 4096)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX8-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 35 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4064 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4064
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4064 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4064
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4064(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[COPY4]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 38 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX6-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068
; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX6-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX6-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.2: ; GFX6-NEXT: bb.2:
; GFX6-NEXT: successors: %bb.3(0x80000000) ; GFX6-NEXT: successors: %bb.3(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY16]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY17]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY17]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4068 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4068
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX7-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068
; GFX7-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX7-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX7-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX7-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.2: ; GFX7-NEXT: bb.2:
; GFX7-NEXT: successors: %bb.3(0x80000000) ; GFX7-NEXT: successors: %bb.3(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4068(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY16]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY17]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY17]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4068 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4068
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX8-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX8-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068 ; GFX8-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4068
; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX8-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX8-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX8-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX8-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.2: ; GFX8-NEXT: bb.2:
; GFX8-NEXT: successors: %bb.3(0x80000000) ; GFX8-NEXT: successors: %bb.3(0x80000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 36 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX6-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX6-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX6-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX6-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX6-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.2: ; GFX6-NEXT: bb.2:
; GFX6-NEXT: successors: %bb.3(0x80000000) ; GFX6-NEXT: successors: %bb.3(0x80000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY16]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY17]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY17]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4096 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4096
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX7-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX7-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX7-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX7-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX7-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.2: ; GFX7-NEXT: bb.2:
; GFX7-NEXT: successors: %bb.3(0x80000000) ; GFX7-NEXT: successors: %bb.3(0x80000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_rsrc_add_4096(<4 x i32> %rsrc, i32 inreg %soffset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY16]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY17]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY17]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4096 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_rsrc_add_4096
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $sgpr2, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr3
; GFX8-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX8-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096 ; GFX8-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 4096
; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def $scc ; GFX8-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY4]], [[S_MOV_B32_]], implicit-def dead $scc
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]] ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX8-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX8-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX8-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec ; GFX8-NEXT: [[S_MOV_B64_:%[0-9]+]]:sreg_64_xexec = S_MOV_B64 $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.2: ; GFX8-NEXT: bb.2:
; GFX8-NEXT: successors: %bb.3(0x80000000) ; GFX8-NEXT: successors: %bb.3(0x80000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_1:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY1]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_2:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY2]], implicit $exec
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY9:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY9]], [[COPY7]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY5]], [[REG_SEQUENCE1]], [[S_MOV_B32_1]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 52 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 12 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 12 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_5000(<4 x i32> %rsrc, i32 %offset.base) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 936, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 936, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 952, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 952, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 52 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 12 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 12 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4076(<4 x i32> %rsrc, i32 %offset.base) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 52 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 12 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 16, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY15]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY16]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY16]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 12 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_add_4080(<4 x i32> %rsrc, i32 %offset.base) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY8:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY8]], [[COPY6]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFEN1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFEN [[COPY4]], [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128), align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 51 Lines • ▼ Show 20 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX6-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX6-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX6-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX6-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX6-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX6-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX6-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: bb.3: ; GFX6-NEXT: bb.3:
; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX6-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX6-NEXT: {{ $}} ; GFX6-NEXT: {{ $}}
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4) ; GFX6-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX6-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
; GFX6-NEXT: $vgpr1 = COPY [[COPY9]] ; GFX6-NEXT: $vgpr1 = COPY [[COPY9]]
; GFX6-NEXT: $vgpr2 = COPY [[COPY10]] ; GFX6-NEXT: $vgpr2 = COPY [[COPY10]]
; GFX6-NEXT: $vgpr3 = COPY [[COPY11]] ; GFX6-NEXT: $vgpr3 = COPY [[COPY11]]
; GFX6-NEXT: $vgpr4 = COPY [[COPY12]] ; GFX6-NEXT: $vgpr4 = COPY [[COPY12]]
; GFX6-NEXT: $vgpr5 = COPY [[COPY13]] ; GFX6-NEXT: $vgpr5 = COPY [[COPY13]]
; GFX6-NEXT: $vgpr6 = COPY [[COPY14]] ; GFX6-NEXT: $vgpr6 = COPY [[COPY14]]
; GFX6-NEXT: $vgpr7 = COPY [[COPY15]] ; GFX6-NEXT: $vgpr7 = COPY [[COPY15]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064 ; GFX7-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: successors: %bb.2(0x80000000) ; GFX7-NEXT: successors: %bb.2(0x80000000)
; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX7-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX7-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX7-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX7-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX7-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX7-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX7-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX7-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: bb.3: ; GFX7-NEXT: bb.3:
; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX7-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4) ; GFX7-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX7-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 18 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
; GFX7-NEXT: $vgpr1 = COPY [[COPY9]] ; GFX7-NEXT: $vgpr1 = COPY [[COPY9]]
; GFX7-NEXT: $vgpr2 = COPY [[COPY10]] ; GFX7-NEXT: $vgpr2 = COPY [[COPY10]]
; GFX7-NEXT: $vgpr3 = COPY [[COPY11]] ; GFX7-NEXT: $vgpr3 = COPY [[COPY11]]
; GFX7-NEXT: $vgpr4 = COPY [[COPY12]] ; GFX7-NEXT: $vgpr4 = COPY [[COPY12]]
; GFX7-NEXT: $vgpr5 = COPY [[COPY13]] ; GFX7-NEXT: $vgpr5 = COPY [[COPY13]]
; GFX7-NEXT: $vgpr6 = COPY [[COPY14]] ; GFX7-NEXT: $vgpr6 = COPY [[COPY14]]
; GFX7-NEXT: $vgpr7 = COPY [[COPY15]] ; GFX7-NEXT: $vgpr7 = COPY [[COPY15]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
;
; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064 ; GFX8-LABEL: name: s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: successors: %bb.2(0x80000000) ; GFX8-NEXT: successors: %bb.2(0x80000000)
; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; GFX8-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX8-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 11 Linesdefine amdgpu_ps <8 x float> @s_buffer_load_v8f32_vgpr_offset_vgpr_rsrc_offset_4064(<4 x i32> %rsrc, i32 %offset.base) {
; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; GFX8-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX8-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX8-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX8-NEXT: [[COPY4:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX8-NEXT: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX8-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX8-NEXT: [[COPY6:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX8-NEXT: [[COPY7:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY6]], [[COPY4]], implicit $exec
; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec ; GFX8-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY7]], [[COPY5]], implicit $exec
; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX8-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX8-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: bb.3: ; GFX8-NEXT: bb.3:
; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX8-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4064, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4) ; GFX8-NEXT: [[BUFFER_LOAD_DWORDX4_OFFSET1:%[0-9]+]]:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET [[REG_SEQUENCE1]], [[S_MOV_B32_]], 4080, 0, 0, implicit $exec :: (dereferenceable invariant load (s128) from unknown-address + 4064, align 4)
; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX8-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 37 Linesdefine amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 18 Linesdefine amdgpu_ps float @s_buffer_load_f32_offset_add_sgpr_vgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX6-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX6-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX6-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[COPY4]], [[REG_SEQUENCE]], [[COPY5]], 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps float @s_buffer_load_f32_offset_add_vgpr_sgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec ; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_vgpr_sgpr_imm
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 25 Linesdefine amdgpu_ps float @s_buffer_load_f32_offset_add_sgpr_vgpr_imm(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec ; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_sgpr_vgpr_imm
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 26 Linesdefine amdgpu_ps float @s_buffer_load_f32_offset_add_imm_sgpr_vgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec ; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY6]], [[COPY4]], 0, implicit $exec
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_imm_sgpr_vgpr
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 25 Linesdefine amdgpu_ps float @s_buffer_load_f32_offset_add_imm_vgpr_sgpr(<4 x i32> inreg %rsrc, i32 %offset.v, i32 inreg %offset.s) {
; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX6-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX6-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX6-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec ; GFX6-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX6-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX6-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX6-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX6-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX7-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr ; GFX7-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr
; GFX7: bb.1 (%ir-block.0): ; GFX7: bb.1 (%ir-block.0):
; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX7-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX7-NEXT: {{ $}} ; GFX7-NEXT: {{ $}}
; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX7-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX7-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX7-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX7-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; GFX7-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX7-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX7-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]] ; GFX7-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[COPY5]]
; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec ; GFX7-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 [[COPY4]], [[COPY6]], 0, implicit $exec
; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; GFX7-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32)) ; GFX7-NEXT: [[BUFFER_LOAD_DWORD_OFFEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_OFFEN [[V_ADD_CO_U32_e64_]], [[REG_SEQUENCE]], [[S_MOV_B32_]], 1024, 0, 0, implicit $exec :: (dereferenceable invariant load (s32))
; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]] ; GFX7-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_DWORD_OFFEN]]
; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; GFX7-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; GFX8-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr ; GFX8-LABEL: name: s_buffer_load_f32_offset_add_imm_vgpr_sgpr
; GFX8: bb.1 (%ir-block.0): ; GFX8: bb.1 (%ir-block.0):
; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; GFX8-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; GFX8-NEXT: {{ $}} ; GFX8-NEXT: {{ $}}
; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX8-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX8-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX8-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX8-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 41 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.atomic.add.ll

Show First 20 Lines • Show All 128 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @struct_buffer_atomic_add_i32__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_BOTHEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_BOTHEN_RTN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_BOTHEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_BOTHEN_RTN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 42 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_i32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: BUFFER_ATOMIC_ADD_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_ADD_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 41 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.atomic.cmpswap.ll

Show First 20 Lines • Show All 84 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @struct_buffer_atomic_cmpswap_i32__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1
; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
▲ Show 20 Lines • Show All 46 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_buffer_atomic_cmpswap_i32_noret__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1
; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_BOTHEN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_BOTHEN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
▲ Show 20 Lines • Show All 45 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.atomic.fadd.ll

Show All 14 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 21 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 20 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__4095_voffset__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__4095_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__4095_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 18 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(float %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 37 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 17 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY11:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY11:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY12:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY12:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; GFX90A-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 38 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 15 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 22 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(float %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset_slc(float %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset_slc ; GFX90A-LABEL: name: struct_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset_slc
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<2 x half> %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @struct_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(<2 x half> %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>), align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 14 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.format.f16.ll

Show All 13 Linesdefine amdgpu_ps half @struct_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 32 Linesdefine amdgpu_ps <2 x half> @struct_buffer_load_format_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY8]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY8]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY11]], [[V_AND_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY11]], [[V_AND_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_buffer_load_format_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_load_format_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 48 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x half> @struct_buffer_load_format_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY10]], [[COPY15]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY10]], [[COPY15]], implicit $exec
; UNPACKED-NEXT: [[COPY16:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY16:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY16]], [[V_AND_B32_e64_3]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY16]], [[V_AND_B32_e64_3]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]] ; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; PACKED-LABEL: name: struct_buffer_load_format_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_load_format_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 40 Linesdefine amdgpu_ps <4 x half> @struct_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 23 Linesdefine amdgpu_ps <4 x half> @struct_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[COPY21:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY21:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY16]], [[COPY21]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY16]], [[COPY21]], implicit $exec
; UNPACKED-NEXT: [[COPY22:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY22:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY22]], [[V_AND_B32_e64_3]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY22]], [[V_AND_B32_e64_3]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]] ; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; PACKED-LABEL: name: struct_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: struct_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 15 Linesdefine amdgpu_ps <4 x half> @struct_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 26 Linesdefine amdgpu_ps half @struct_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffsset_add_4095(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset.base, i32 inreg %soffset) {
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffsset_add_4095 ; PACKED-LABEL: name: struct_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffsset_add_4095
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 22 Linesdefine amdgpu_ps half @struct_buffer_load_format_i16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_buffer_load_format_i16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_load_format_i16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.format.ll

Show First 20 Lines • Show All 127 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x float> @struct_buffer_load_format_v4f32__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 187 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.load.ll

Show First 20 Lines • Show All 198 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @struct_buffer_load_f32__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 215 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.store.format.f16.ll

Show All 13 Linesdefine amdgpu_ps void @struct_buffer_store_format_f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(half %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_buffer_store_format_f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_store_format_f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 25 Linesdefine amdgpu_ps void @struct_buffer_store_format_v2f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<2 x half> %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_BOTHEN_exact [[REG_SEQUENCE1]], [[REG_SEQUENCE2]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_BOTHEN_exact [[REG_SEQUENCE1]], [[REG_SEQUENCE2]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_buffer_store_format_v2f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_store_format_v2f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 34 Linesdefine amdgpu_ps void @struct_buffer_store_format_v4f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x half> %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY10]], [[COPY1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY10]], [[COPY1]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY7]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY7]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_BOTHEN_exact [[REG_SEQUENCE1]], [[REG_SEQUENCE2]], [[REG_SEQUENCE]], [[COPY8]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_BOTHEN_exact [[REG_SEQUENCE1]], [[REG_SEQUENCE2]], [[REG_SEQUENCE]], [[COPY8]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_buffer_store_format_v4f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_store_format_v4f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
Show All 40 Linesdefine amdgpu_ps void @struct_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(half inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: struct_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 17 Linesdefine amdgpu_ps void @struct_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(half inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 22 Linesdefine amdgpu_ps void @struct_buffer_store_format_i16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(i16 %val, <4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16), align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_buffer_store_format_i16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_buffer_store_format_i16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.store.format.f32.ll

Show First 20 Lines • Show All 122 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_buffer_store_format_f32__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(float inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: BUFFER_STORE_FORMAT_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_FORMAT_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 39 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.buffer.store.ll

Show First 20 Lines • Show All 127 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_buffer_store_v4f32_vgpr_rsrc__sgpr_val__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x float> inreg %val, <4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY15:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY17:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY17:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY17]], [[COPY15]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY17]], [[COPY15]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY12]], %subreg.sub0, [[COPY13]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY12]], %subreg.sub0, [[COPY13]], %subreg.sub1
; CHECK-NEXT: BUFFER_STORE_DWORDX4_BOTHEN_exact [[COPY11]], [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>), align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORDX4_BOTHEN_exact [[COPY11]], [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 140 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.atomic.add.ll

Show First 20 Lines • Show All 128 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @struct_ptr_buffer_atomic_add_i32__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_BOTHEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_BOTHEN_RTN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_ADD_BOTHEN_RTN:%[0-9]+]]:vgpr_32 = BUFFER_ATOMIC_ADD_BOTHEN_RTN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 42 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_i32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: BUFFER_ATOMIC_ADD_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_ADD_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 41 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.atomic.cmpswap.ll

Show First 20 Lines • Show All 84 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @struct_ptr_buffer_atomic_cmpswap_i32__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1
; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN:%[0-9]+]]:vreg_64 = BUFFER_ATOMIC_CMPSWAP_BOTHEN_RTN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 1, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
▲ Show 20 Lines • Show All 46 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_cmpswap_i32_noret__sgpr_val__sgpr_cmp__vgpr_rsrc__sgpr_voffset__vgpr_soffset(i32 inreg %val, i32 inreg %cmp, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY5]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY15]], [[COPY13]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY8]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY8]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY11]], %subreg.sub0, [[COPY12]], %subreg.sub1
; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_BOTHEN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_ATOMIC_CMPSWAP_BOTHEN [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
▲ Show 20 Lines • Show All 45 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.atomic.fadd.ll

Show All 14 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 21 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset_plus4095__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 20 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__4095_voffset__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__4095_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__4095_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 18 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(float %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 37 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 17 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__sgpr_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY11:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY11:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY12:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY12:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; GFX90A-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64_align2 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 38 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 %soffset) {
; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX908-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX908-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX908-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX908-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX908-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX908-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX908-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.3: ; GFX908-NEXT: bb.3:
; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX908-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX908-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX908-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.4: ; GFX908-NEXT: bb.4:
; GFX908-NEXT: successors: %bb.5(0x80000000) ; GFX908-NEXT: successors: %bb.5(0x80000000)
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; GFX908-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; GFX908-NEXT: {{ $}} ; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.5: ; GFX908-NEXT: bb.5:
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: successors: %bb.2(0x80000000) ; GFX90A-NEXT: successors: %bb.2(0x80000000)
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 15 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__sgpr_val__vgpr_rsrc__0_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 %soffset) {
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; GFX90A-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1 ; GFX90A-NEXT: [[COPY9:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub0_sub1
; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3 ; GFX90A-NEXT: [[COPY10:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]].sub2_sub3
; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; GFX90A-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; GFX90A-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; GFX90A-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; GFX90A-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; GFX90A-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: bb.3: ; GFX90A-NEXT: bb.3:
; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; GFX90A-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX90A-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY7]], [[COPY8]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; GFX90A-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; GFX90A-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
Show All 22 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc(float %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset_slc
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset_slc(float %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_ADD_F32_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 2, implicit $exec :: (volatile dereferenceable load store (s32) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset_slc ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_f32_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset_slc
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset(<2 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; GFX908-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_BOTHEN [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__vgpr_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 19 Linesdefine amdgpu_ps void @struct_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset(<2 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 inreg %soffset) {
; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX908-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX908-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; GFX908-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; GFX908-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; GFX908-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; GFX908-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8) ; GFX908-NEXT: BUFFER_ATOMIC_PK_ADD_F16_IDXEN [[COPY]], [[COPY5]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, implicit $exec :: (volatile dereferenceable load store (<2 x s16>) on %ir.rsrc, align 1, addrspace 8)
; GFX908-NEXT: S_ENDPGM 0 ; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset ; GFX90A-LABEL: name: struct_ptr_buffer_atomic_add_v2f16_noret__vgpr_val__sgpr_rsrc__0_voffset__sgpr_soffset
; GFX90A: bb.1 (%ir-block.0): ; GFX90A: bb.1 (%ir-block.0):
; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; GFX90A-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; GFX90A-NEXT: {{ $}} ; GFX90A-NEXT: {{ $}}
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; GFX90A-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; GFX90A-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; GFX90A-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; GFX90A-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 14 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.load.format.f16.ll

Show All 13 Linesdefine amdgpu_ps half @struct_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 32 Linesdefine amdgpu_ps <2 x half> @struct_ptr_buffer_load_format_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY8]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY8]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY11]], [[V_AND_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY11]], [[V_AND_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_]], [[V_LSHLREV_B32_e64_]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_ptr_buffer_load_format_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_load_format_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 48 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x half> @struct_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY15:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY10]], [[COPY15]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY10]], [[COPY15]], implicit $exec
; UNPACKED-NEXT: [[COPY16:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY16:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY16]], [[V_AND_B32_e64_3]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY16]], [[V_AND_B32_e64_3]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]] ; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; PACKED-LABEL: name: struct_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_load_format_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 40 Linesdefine amdgpu_ps <4 x half> @struct_ptr_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 23 Linesdefine amdgpu_ps <4 x half> @struct_ptr_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[COPY21:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY21:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY16]], [[COPY21]], implicit $exec ; UNPACKED-NEXT: [[V_AND_B32_e64_3:%[0-9]+]]:vgpr_32 = V_AND_B32_e64 [[COPY16]], [[COPY21]], implicit $exec
; UNPACKED-NEXT: [[COPY22:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]] ; UNPACKED-NEXT: [[COPY22:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_1]]
; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY22]], [[V_AND_B32_e64_3]], implicit $exec ; UNPACKED-NEXT: [[V_LSHLREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHLREV_B32_e64 [[COPY22]], [[V_AND_B32_e64_3]], implicit $exec
; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec ; UNPACKED-NEXT: [[V_OR_B32_e64_1:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[V_AND_B32_e64_2]], [[V_LSHLREV_B32_e64_1]], implicit $exec
; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]] ; UNPACKED-NEXT: $vgpr0 = COPY [[V_OR_B32_e64_]]
; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]] ; UNPACKED-NEXT: $vgpr1 = COPY [[V_OR_B32_e64_1]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; PACKED-LABEL: name: struct_ptr_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 15 Linesdefine amdgpu_ps <4 x half> @struct_ptr_buffer_load_format_v4f16__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = BUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 26 Linesdefine amdgpu_ps half @struct_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffsset_add_4095(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset.base, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffsset_add_4095 ; PACKED-LABEL: name: struct_ptr_buffer_load_format_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffsset_add_4095
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 22 Linesdefine amdgpu_ps half @struct_ptr_buffer_load_format_i16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]] ; UNPACKED-NEXT: $vgpr0 = COPY [[BUFFER_LOAD_FORMAT_D16_X_gfx80_BOTHEN]]
; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; UNPACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; PACKED-LABEL: name: struct_ptr_buffer_load_format_i16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_load_format_i16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.load.format.ll

Show First 20 Lines • Show All 127 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x float> @struct_ptr_buffer_load_format_v4f32__vpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = BUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 187 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.load.ll

Show First 20 Lines • Show All 198 Lines • ▼ Show 20 Linesdefine amdgpu_ps float @struct_ptr_buffer_load_f32__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; CHECK-NEXT: [[BUFFER_LOAD_DWORD_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[BUFFER_LOAD_DWORD_BOTHEN:%[0-9]+]]:vgpr_32 = BUFFER_LOAD_DWORD_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable load (s32) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 215 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.store.format.f16.ll

Show All 13 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(half %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_ptr_buffer_store_format_f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_store_format_f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 25 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_v2f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<2 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16 ; UNPACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 16
; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY8]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_BOTHEN_exact [[REG_SEQUENCE]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XY_gfx80_BOTHEN_exact [[REG_SEQUENCE]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (<2 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_ptr_buffer_store_format_v2f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_store_format_v2f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 34 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_v4f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x half> %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY9]], [[COPY]], implicit $exec
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY10]], [[COPY1]], implicit $exec ; UNPACKED-NEXT: [[V_LSHRREV_B32_e64_1:%[0-9]+]]:vgpr_32 = V_LSHRREV_B32_e64 [[COPY10]], [[COPY1]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[V_LSHRREV_B32_e64_]], %subreg.sub1, [[COPY1]], %subreg.sub2, [[V_LSHRREV_B32_e64_1]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1, [[COPY4]], %subreg.sub2, [[COPY5]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY3]], %subreg.sub1, [[COPY4]], %subreg.sub2, [[COPY5]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY7]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY7]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_BOTHEN_exact [[REG_SEQUENCE]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[COPY8]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_XYZW_gfx80_BOTHEN_exact [[REG_SEQUENCE]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[COPY8]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s16>) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_ptr_buffer_store_format_v4f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_store_format_v4f16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2, $vgpr3 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2, $vgpr3
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr2
Show All 40 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(half inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; UNPACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.4: ; UNPACKED-NEXT: bb.4:
; UNPACKED-NEXT: successors: %bb.5(0x80000000) ; UNPACKED-NEXT: successors: %bb.5(0x80000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; UNPACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.5: ; UNPACKED-NEXT: bb.5:
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_ptr_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: successors: %bb.2(0x80000000) ; PACKED-NEXT: successors: %bb.2(0x80000000)
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr1
Show All 17 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_f16__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(half inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 22 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_i16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(i16 %val, ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY4:%[0-9]+]]:sreg_32 = COPY $sgpr5
; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; UNPACKED-NEXT: [[COPY7:%[0-9]+]]:sreg_32 = COPY $sgpr6
; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY1]], %subreg.sub0, [[COPY2]], %subreg.sub1, [[COPY3]], %subreg.sub2, [[COPY4]], %subreg.sub3
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY5]], %subreg.sub0, [[COPY6]], %subreg.sub1
; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: BUFFER_STORE_FORMAT_D16_X_gfx80_BOTHEN_exact [[COPY]], [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY7]], 0, 0, 0, implicit $exec :: (dereferenceable store (s16) into %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: S_ENDPGM 0 ; UNPACKED-NEXT: S_ENDPGM 0
;
; PACKED-LABEL: name: struct_ptr_buffer_store_format_i16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; PACKED-LABEL: name: struct_ptr_buffer_store_format_i16__vgpr_val__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; PACKED: bb.1 (%ir-block.0): ; PACKED: bb.1 (%ir-block.0):
; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2 ; PACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1, $vgpr2
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; PACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr2
; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; PACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr3
; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; PACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr4
Show All 17 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.store.format.f32.ll

Show First 20 Lines • Show All 122 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_format_f32__sgpr_val__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(float inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY4]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY13:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY13]], [[COPY11]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY14]], [[COPY12]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY7]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY9]], %subreg.sub0, [[COPY10]], %subreg.sub1
; CHECK-NEXT: BUFFER_STORE_FORMAT_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_FORMAT_X_BOTHEN_exact [[COPY8]], [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (s32) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
Show All 39 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.buffer.store.ll

Show First 20 Lines • Show All 127 Lines • ▼ Show 20 Linesdefine amdgpu_ps void @struct_ptr_buffer_store_v4f32_vgpr_rsrc__sgpr_val__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x float> inreg %val, ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY7]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY14:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY15:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY15:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1 ; CHECK-NEXT: [[COPY16:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub0_sub1
; CHECK-NEXT: [[COPY17:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3 ; CHECK-NEXT: [[COPY17:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE2]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY16]], [[COPY14]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY17]], [[COPY15]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY17]], [[COPY15]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY10]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY12]], %subreg.sub0, [[COPY13]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE3:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY12]], %subreg.sub0, [[COPY13]], %subreg.sub1
; CHECK-NEXT: BUFFER_STORE_DWORDX4_BOTHEN_exact [[COPY11]], [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>) into %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: BUFFER_STORE_DWORDX4_BOTHEN_exact [[COPY11]], [[REG_SEQUENCE3]], [[REG_SEQUENCE2]], [[V_READFIRSTLANE_B32_4]], 0, 0, 0, implicit $exec :: (dereferenceable store (<4 x s32>) into %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 140 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.tbuffer.load.f16.ll

Show All 15 Linesdefine amdgpu_ps half @struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps <2 x half> @struct_tbuffer_load_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 41 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x half> @struct_tbuffer_load_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY7]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY7]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY8]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY8]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 43 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_vindex0(ptr addrspace(8) inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; PACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY4]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY4]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_vindex0 ; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_vindex0
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 37 Linesdefine amdgpu_ps <4 x half> @struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.4: ; PACKED-NEXT: bb.4:
; PACKED-NEXT: successors: %bb.5(0x80000000) ; PACKED-NEXT: successors: %bb.5(0x80000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.5: ; PACKED-NEXT: bb.5:
; PACKED-NEXT: [[COPY13:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0 ; PACKED-NEXT: [[COPY13:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0
; PACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1 ; PACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY13]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY13]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY14]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY14]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: successors: %bb.2(0x80000000) ; UNPACKED-NEXT: successors: %bb.2(0x80000000)
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 15 Linesdefine amdgpu_ps <4 x half> @struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>) from %ir.rsrc, align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 44 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffset_add4095(ptr addrspace(8) inreg %rsrc, i32 %vindex, i32 %voffset.base, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 78, 0, 0, implicit $exec :: (dereferenceable load (s16) from %ir.rsrc, align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffset_add4095 ; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffset_add4095
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.ptr.tbuffer.load.ll

Show First 20 Lines • Show All 150 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x float> @struct_tbuffer_load_v4f32__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(ptr addrspace(8) %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>) from %ir.rsrc, align 1, addrspace 8) ; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>) from %ir.rsrc, align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 49 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.tbuffer.load.f16.ll

Show All 15 Linesdefine amdgpu_ps half @struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 21 Linesdefine amdgpu_ps <2 x half> @struct_tbuffer_load_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<2 x s16>), align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_XY_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_v2f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 41 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x half> @struct_tbuffer_load_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0 ; PACKED-NEXT: [[COPY7:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0
; PACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1 ; PACKED-NEXT: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY7]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY7]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY8]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY8]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
▲ Show 20 Lines • Show All 43 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_vindex0(<4 x i32> inreg %rsrc, i32 %voffset, i32 inreg %soffset) {
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 ; PACKED-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0
; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] ; PACKED-NEXT: [[COPY6:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY4]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY6]], %subreg.sub0, [[COPY4]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY5]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_vindex0 ; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_vindex0
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 37 Linesdefine amdgpu_ps <4 x half> @struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; PACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; PACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; PACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; PACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; PACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; PACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; PACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.3: ; PACKED-NEXT: bb.3:
; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; PACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN:%[0-9]+]]:vreg_64 = TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; PACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec ; PACKED-NEXT: SI_WATERFALL_LOOP %bb.2, implicit $exec
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.4: ; PACKED-NEXT: bb.4:
; PACKED-NEXT: successors: %bb.5(0x80000000) ; PACKED-NEXT: successors: %bb.5(0x80000000)
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]] ; PACKED-NEXT: $exec = S_MOV_B64_term [[S_MOV_B64_]]
; PACKED-NEXT: {{ $}} ; PACKED-NEXT: {{ $}}
; PACKED-NEXT: bb.5: ; PACKED-NEXT: bb.5:
; PACKED-NEXT: [[COPY13:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0 ; PACKED-NEXT: [[COPY13:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub0
; PACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1 ; PACKED-NEXT: [[COPY14:%[0-9]+]]:vgpr_32 = COPY [[TBUFFER_LOAD_FORMAT_D16_XYZW_BOTHEN]].sub1
; PACKED-NEXT: $vgpr0 = COPY [[COPY13]] ; PACKED-NEXT: $vgpr0 = COPY [[COPY13]]
; PACKED-NEXT: $vgpr1 = COPY [[COPY14]] ; PACKED-NEXT: $vgpr1 = COPY [[COPY14]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0, implicit $vgpr1
;
; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset ; UNPACKED-LABEL: name: struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: successors: %bb.2(0x80000000) ; UNPACKED-NEXT: successors: %bb.2(0x80000000)
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
Show All 15 Linesdefine amdgpu_ps <4 x half> @struct_tbuffer_load_v4f16__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; UNPACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; UNPACKED-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; UNPACKED-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; UNPACKED-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; UNPACKED-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; UNPACKED-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; UNPACKED-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; UNPACKED-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: bb.3: ; UNPACKED-NEXT: bb.3:
; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; UNPACKED-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; UNPACKED-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8) ; UNPACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_D16_XYZW_gfx80_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s16>), align 1, addrspace 8)
; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; UNPACKED-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 44 Lines • ▼ Show 20 Linesdefine amdgpu_ps half @struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffset_add4095(<4 x i32> inreg %rsrc, i32 %vindex, i32 %voffset.base, i32 inreg %soffset) {
; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3 ; PACKED-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[COPY]], %subreg.sub0, [[COPY1]], %subreg.sub1, [[COPY2]], %subreg.sub2, [[COPY3]], %subreg.sub3
; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0 ; PACKED-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1 ; PACKED-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6 ; PACKED-NEXT: [[COPY6:%[0-9]+]]:sreg_32 = COPY $sgpr6
; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1 ; PACKED-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY5]], %subreg.sub1
; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8) ; PACKED-NEXT: [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN:%[0-9]+]]:vgpr_32 = TBUFFER_LOAD_FORMAT_D16_X_BOTHEN [[REG_SEQUENCE1]], [[REG_SEQUENCE]], [[COPY6]], 4095, 78, 0, 0, implicit $exec :: (dereferenceable load (s16), align 1, addrspace 8)
; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]] ; PACKED-NEXT: $vgpr0 = COPY [[TBUFFER_LOAD_FORMAT_D16_X_BOTHEN]]
; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0 ; PACKED-NEXT: SI_RETURN_TO_EPILOG implicit $vgpr0
;
; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffset_add4095 ; UNPACKED-LABEL: name: struct_tbuffer_load_f16__sgpr_rsrc__vgpr_vindex__vgpr_voffset__sgpr_soffset_voffset_add4095
; UNPACKED: bb.1 (%ir-block.0): ; UNPACKED: bb.1 (%ir-block.0):
; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1 ; UNPACKED-NEXT: liveins: $sgpr2, $sgpr3, $sgpr4, $sgpr5, $sgpr6, $vgpr0, $vgpr1
; UNPACKED-NEXT: {{ $}} ; UNPACKED-NEXT: {{ $}}
; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2 ; UNPACKED-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr2
; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3 ; UNPACKED-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr3
; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4 ; UNPACKED-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY $sgpr4
; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5 ; UNPACKED-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY $sgpr5
Show All 19 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.struct.tbuffer.load.ll

Show First 20 Lines • Show All 150 Lines • ▼ Show 20 Linesdefine amdgpu_ps <4 x float> @struct_tbuffer_load_v4f32__vgpr_rsrc__sgpr_vindex__sgpr_voffset__vgpr_soffset(<4 x i32> %rsrc, i32 inreg %vindex, i32 inreg %voffset, i32 %soffset) {
; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_3:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY3]], implicit $exec
; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3 ; CHECK-NEXT: [[REG_SEQUENCE1:%[0-9]+]]:sgpr_128 = REG_SEQUENCE [[V_READFIRSTLANE_B32_]], %subreg.sub0, [[V_READFIRSTLANE_B32_1]], %subreg.sub1, [[V_READFIRSTLANE_B32_2]], %subreg.sub2, [[V_READFIRSTLANE_B32_3]], %subreg.sub3
; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1 ; CHECK-NEXT: [[COPY9:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub0_sub1
; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3 ; CHECK-NEXT: [[COPY10:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]].sub2_sub3
; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1 ; CHECK-NEXT: [[COPY11:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub0_sub1
; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3 ; CHECK-NEXT: [[COPY12:%[0-9]+]]:sreg_64 = COPY [[REG_SEQUENCE1]].sub2_sub3
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY11]], [[COPY9]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U64_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U64_e64 [[COPY12]], [[COPY10]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[V_CMP_EQ_U64_e64_]], [[V_CMP_EQ_U64_e64_1]], implicit-def dead $scc
; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_READFIRSTLANE_B32_4:%[0-9]+]]:sreg_32 = V_READFIRSTLANE_B32 [[COPY6]], implicit $exec
; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec ; CHECK-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[V_READFIRSTLANE_B32_4]], [[COPY6]], implicit $exec
; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def $scc ; CHECK-NEXT: [[S_AND_B64_1:%[0-9]+]]:sreg_64_xexec = S_AND_B64 [[S_AND_B64_]], [[V_CMP_EQ_U32_e64_]], implicit-def dead $scc
; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec ; CHECK-NEXT: [[S_AND_SAVEEXEC_B64_:%[0-9]+]]:sreg_64_xexec = S_AND_SAVEEXEC_B64 killed [[S_AND_B64_1]], implicit-def $exec, implicit-def $scc, implicit $exec
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: bb.3: ; CHECK-NEXT: bb.3:
; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000) ; CHECK-NEXT: successors: %bb.4(0x40000000), %bb.2(0x40000000)
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1 ; CHECK-NEXT: [[REG_SEQUENCE2:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY7]], %subreg.sub0, [[COPY8]], %subreg.sub1
; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>), align 1, addrspace 8) ; CHECK-NEXT: [[TBUFFER_LOAD_FORMAT_XYZW_BOTHEN:%[0-9]+]]:vreg_128 = TBUFFER_LOAD_FORMAT_XYZW_BOTHEN [[REG_SEQUENCE2]], [[REG_SEQUENCE1]], [[V_READFIRSTLANE_B32_4]], 0, 78, 0, 0, implicit $exec :: (dereferenceable load (<4 x s32>), align 1, addrspace 8)
; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc ; CHECK-NEXT: $exec = S_XOR_B64_term $exec, [[S_AND_SAVEEXEC_B64_]], implicit-def $scc
▲ Show 20 Lines • Show All 49 Lines • Show Last 20 Lines

llvm/test/CodeGen/Mips/GlobalISel/instruction-select/branch.mir

Show All 27 Lines
name: Unconditional_branch name: Unconditional_branch
alignment: 4 alignment: 4
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
; MIPS32-LABEL: name: Unconditional_branch ; MIPS32-LABEL: name: Unconditional_branch
; MIPS32: bb.0.entry: ; MIPS32: bb.0.entry:
; MIPS32: successors: %bb.2(0x80000000) ; MIPS32-NEXT: successors: %bb.2(0x80000000)
; MIPS32: liveins: $a0, $a1 ; MIPS32-NEXT: liveins: $a0, $a1
; MIPS32: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32-NEXT: {{ $}}
; MIPS32: J %bb.2, implicit-def $at ; MIPS32-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32: bb.1.end: ; MIPS32-NEXT: J %bb.2, implicit-def dead $at
; MIPS32: $v0 = COPY [[COPY]] ; MIPS32-NEXT: {{ $}}
; MIPS32: RetRA implicit $v0 ; MIPS32-NEXT: bb.1.end:
; MIPS32: bb.2.block: ; MIPS32-NEXT: $v0 = COPY [[COPY]]
; MIPS32: successors: %bb.1(0x80000000) ; MIPS32-NEXT: RetRA implicit $v0
; MIPS32: J %bb.1, implicit-def $at ; MIPS32-NEXT: {{ $}}
; MIPS32-NEXT: bb.2.block:
; MIPS32-NEXT: successors: %bb.1(0x80000000)
; MIPS32-NEXT: {{ $}}
; MIPS32-NEXT: J %bb.1, implicit-def dead $at
bb.1.entry: bb.1.entry:
liveins: $a0, $a1 liveins: $a0, $a1
%0:gprb(s32) = COPY $a0 %0:gprb(s32) = COPY $a0
G_BR %bb.3 G_BR %bb.3
bb.2.end: bb.2.end:
$v0 = COPY %0(s32) $v0 = COPY %0(s32)
RetRA implicit $v0 RetRA implicit $v0
bb.3.block: bb.3.block:
G_BR %bb.2 G_BR %bb.2
... ...
--- ---
name: Conditional_branch name: Conditional_branch
alignment: 4 alignment: 4
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
; MIPS32-LABEL: name: Conditional_branch ; MIPS32-LABEL: name: Conditional_branch
; MIPS32: bb.0 (%ir-block.0): ; MIPS32: bb.0 (%ir-block.0):
; MIPS32: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32: liveins: $a0, $a1, $a2 ; MIPS32-NEXT: liveins: $a0, $a1, $a2
; MIPS32: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32-NEXT: {{ $}}
; MIPS32: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1 ; MIPS32-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1
; MIPS32: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32: J %bb.2, implicit-def $at ; MIPS32-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32: bb.1.if.then: ; MIPS32-NEXT: J %bb.2, implicit-def dead $at
; MIPS32: $v0 = COPY [[COPY1]] ; MIPS32-NEXT: {{ $}}
; MIPS32: RetRA implicit $v0 ; MIPS32-NEXT: bb.1.if.then:
; MIPS32: bb.2.if.else: ; MIPS32-NEXT: $v0 = COPY [[COPY1]]
; MIPS32: $v0 = COPY [[COPY2]] ; MIPS32-NEXT: RetRA implicit $v0
; MIPS32: RetRA implicit $v0 ; MIPS32-NEXT: {{ $}}
; MIPS32-NEXT: bb.2.if.else:
; MIPS32-NEXT: $v0 = COPY [[COPY2]]
; MIPS32-NEXT: RetRA implicit $v0
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $a0, $a1, $a2 liveins: $a0, $a1, $a2
%3:gprb(s32) = COPY $a0 %3:gprb(s32) = COPY $a0
%1:gprb(s32) = COPY $a1 %1:gprb(s32) = COPY $a1
%2:gprb(s32) = COPY $a2 %2:gprb(s32) = COPY $a2
%5:gprb(s32) = G_CONSTANT i32 1 %5:gprb(s32) = G_CONSTANT i32 1
%6:gprb(s32) = COPY %3(s32) %6:gprb(s32) = COPY %3(s32)
Show All 13 Lines

llvm/test/CodeGen/Mips/GlobalISel/instruction-select/phi.mir

Show First 20 Lines • Show All 67 Lines • ▼ Show 20 Lines
name: phi_i32 name: phi_i32
alignment: 4 alignment: 4
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
; MIPS32FP32-LABEL: name: phi_i32 ; MIPS32FP32-LABEL: name: phi_i32
; MIPS32FP32: bb.0.entry: ; MIPS32FP32: bb.0.entry:
; MIPS32FP32: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP32-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP32: liveins: $a0, $a1, $a2 ; MIPS32FP32-NEXT: liveins: $a0, $a1, $a2
; MIPS32FP32: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1 ; MIPS32FP32-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32FP32: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32FP32-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1
; MIPS32FP32: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32FP32-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32FP32: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP32-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32FP32: J %bb.2, implicit-def $at ; MIPS32FP32-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP32: bb.1.cond.true: ; MIPS32FP32-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: J %bb.3, implicit-def $at ; MIPS32FP32-NEXT: bb.1.cond.true:
; MIPS32FP32: bb.2.cond.false: ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: bb.3.cond.end: ; MIPS32FP32-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP32: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: $v0 = COPY [[PHI]] ; MIPS32FP32-NEXT: bb.2.cond.false:
; MIPS32FP32: RetRA implicit $v0 ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: bb.3.cond.end:
; MIPS32FP32-NEXT: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2
; MIPS32FP32-NEXT: $v0 = COPY [[PHI]]
; MIPS32FP32-NEXT: RetRA implicit $v0
;
; MIPS32FP64-LABEL: name: phi_i32 ; MIPS32FP64-LABEL: name: phi_i32
; MIPS32FP64: bb.0.entry: ; MIPS32FP64: bb.0.entry:
; MIPS32FP64: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP64-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP64: liveins: $a0, $a1, $a2 ; MIPS32FP64-NEXT: liveins: $a0, $a1, $a2
; MIPS32FP64: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1 ; MIPS32FP64-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32FP64: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32FP64-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1
; MIPS32FP64: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32FP64-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32FP64: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP64-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32FP64: J %bb.2, implicit-def $at ; MIPS32FP64-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP64: bb.1.cond.true: ; MIPS32FP64-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: J %bb.3, implicit-def $at ; MIPS32FP64-NEXT: bb.1.cond.true:
; MIPS32FP64: bb.2.cond.false: ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: bb.3.cond.end: ; MIPS32FP64-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP64: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: $v0 = COPY [[PHI]] ; MIPS32FP64-NEXT: bb.2.cond.false:
; MIPS32FP64: RetRA implicit $v0 ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: bb.3.cond.end:
; MIPS32FP64-NEXT: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2
; MIPS32FP64-NEXT: $v0 = COPY [[PHI]]
; MIPS32FP64-NEXT: RetRA implicit $v0
bb.1.entry: bb.1.entry:
liveins: $a0, $a1, $a2 liveins: $a0, $a1, $a2
%3:gprb(s32) = COPY $a0 %3:gprb(s32) = COPY $a0
%1:gprb(s32) = COPY $a1 %1:gprb(s32) = COPY $a1
%2:gprb(s32) = COPY $a2 %2:gprb(s32) = COPY $a2
%6:gprb(s32) = G_CONSTANT i32 1 %6:gprb(s32) = G_CONSTANT i32 1
%7:gprb(s32) = COPY %3(s32) %7:gprb(s32) = COPY %3(s32)
Show All 19 Lines
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
fixedStack: fixedStack:
- { id: 0, offset: 20, size: 4, alignment: 4, isImmutable: true } - { id: 0, offset: 20, size: 4, alignment: 4, isImmutable: true }
- { id: 1, offset: 16, size: 4, alignment: 8, isImmutable: true } - { id: 1, offset: 16, size: 4, alignment: 8, isImmutable: true }
body: | body: |
; MIPS32FP32-LABEL: name: phi_i64 ; MIPS32FP32-LABEL: name: phi_i64
; MIPS32FP32: bb.0.entry: ; MIPS32FP32: bb.0.entry:
; MIPS32FP32: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP32-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP32: liveins: $a0, $a2, $a3 ; MIPS32FP32-NEXT: liveins: $a0, $a2, $a3
; MIPS32FP32: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: [[COPY1:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32FP32-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32FP32: [[COPY2:%[0-9]+]]:gpr32 = COPY $a3 ; MIPS32FP32-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32FP32: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0 ; MIPS32FP32-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a3
; MIPS32FP32: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8) ; MIPS32FP32-NEXT: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0
; MIPS32FP32: [[ADDiu1:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.1, 0 ; MIPS32FP32-NEXT: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8)
; MIPS32FP32: [[LW1:%[0-9]+]]:gpr32 = LW [[ADDiu1]], 0 :: (load (s32) from %fixed-stack.1) ; MIPS32FP32-NEXT: [[ADDiu1:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.1, 0
; MIPS32FP32: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32FP32-NEXT: [[LW1:%[0-9]+]]:gpr32 = LW [[ADDiu1]], 0 :: (load (s32) from %fixed-stack.1)
; MIPS32FP32: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP32-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32FP32: J %bb.2, implicit-def $at ; MIPS32FP32-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP32: bb.1.cond.true: ; MIPS32FP32-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: J %bb.3, implicit-def $at ; MIPS32FP32-NEXT: bb.1.cond.true:
; MIPS32FP32: bb.2.cond.false: ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: bb.3.cond.end: ; MIPS32FP32-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP32: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[LW]], %bb.2 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: [[PHI1:%[0-9]+]]:gpr32 = PHI [[COPY2]], %bb.1, [[LW1]], %bb.2 ; MIPS32FP32-NEXT: bb.2.cond.false:
; MIPS32FP32: $v0 = COPY [[PHI]] ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32: $v1 = COPY [[PHI1]] ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: RetRA implicit $v0, implicit $v1 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: bb.3.cond.end:
; MIPS32FP32-NEXT: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[LW]], %bb.2
; MIPS32FP32-NEXT: [[PHI1:%[0-9]+]]:gpr32 = PHI [[COPY2]], %bb.1, [[LW1]], %bb.2
; MIPS32FP32-NEXT: $v0 = COPY [[PHI]]
; MIPS32FP32-NEXT: $v1 = COPY [[PHI1]]
; MIPS32FP32-NEXT: RetRA implicit $v0, implicit $v1
;
; MIPS32FP64-LABEL: name: phi_i64 ; MIPS32FP64-LABEL: name: phi_i64
; MIPS32FP64: bb.0.entry: ; MIPS32FP64: bb.0.entry:
; MIPS32FP64: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP64-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP64: liveins: $a0, $a2, $a3 ; MIPS32FP64-NEXT: liveins: $a0, $a2, $a3
; MIPS32FP64: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: [[COPY1:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32FP64-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32FP64: [[COPY2:%[0-9]+]]:gpr32 = COPY $a3 ; MIPS32FP64-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32FP64: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0 ; MIPS32FP64-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a3
; MIPS32FP64: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8) ; MIPS32FP64-NEXT: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0
; MIPS32FP64: [[ADDiu1:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.1, 0 ; MIPS32FP64-NEXT: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8)
; MIPS32FP64: [[LW1:%[0-9]+]]:gpr32 = LW [[ADDiu1]], 0 :: (load (s32) from %fixed-stack.1) ; MIPS32FP64-NEXT: [[ADDiu1:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.1, 0
; MIPS32FP64: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32FP64-NEXT: [[LW1:%[0-9]+]]:gpr32 = LW [[ADDiu1]], 0 :: (load (s32) from %fixed-stack.1)
; MIPS32FP64: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP64-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32FP64: J %bb.2, implicit-def $at ; MIPS32FP64-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP64: bb.1.cond.true: ; MIPS32FP64-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: J %bb.3, implicit-def $at ; MIPS32FP64-NEXT: bb.1.cond.true:
; MIPS32FP64: bb.2.cond.false: ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: bb.3.cond.end: ; MIPS32FP64-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP64: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[LW]], %bb.2 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: [[PHI1:%[0-9]+]]:gpr32 = PHI [[COPY2]], %bb.1, [[LW1]], %bb.2 ; MIPS32FP64-NEXT: bb.2.cond.false:
; MIPS32FP64: $v0 = COPY [[PHI]] ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64: $v1 = COPY [[PHI1]] ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: RetRA implicit $v0, implicit $v1 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: bb.3.cond.end:
; MIPS32FP64-NEXT: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[LW]], %bb.2
; MIPS32FP64-NEXT: [[PHI1:%[0-9]+]]:gpr32 = PHI [[COPY2]], %bb.1, [[LW1]], %bb.2
; MIPS32FP64-NEXT: $v0 = COPY [[PHI]]
; MIPS32FP64-NEXT: $v1 = COPY [[PHI1]]
; MIPS32FP64-NEXT: RetRA implicit $v0, implicit $v1
bb.1.entry: bb.1.entry:
liveins: $a0, $a2, $a3 liveins: $a0, $a2, $a3
%3:gprb(s32) = COPY $a0 %3:gprb(s32) = COPY $a0
%4:gprb(s32) = COPY $a2 %4:gprb(s32) = COPY $a2
%5:gprb(s32) = COPY $a3 %5:gprb(s32) = COPY $a3
%8:gprb(p0) = G_FRAME_INDEX %fixed-stack.1 %8:gprb(p0) = G_FRAME_INDEX %fixed-stack.1
%6:gprb(s32) = G_LOAD %8(p0) :: (load (s32) from %fixed-stack.1, align 8) %6:gprb(s32) = G_LOAD %8(p0) :: (load (s32) from %fixed-stack.1, align 8)
Show All 22 Lines
name: phi_float name: phi_float
alignment: 4 alignment: 4
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
; MIPS32FP32-LABEL: name: phi_float ; MIPS32FP32-LABEL: name: phi_float
; MIPS32FP32: bb.0.entry: ; MIPS32FP32: bb.0.entry:
; MIPS32FP32: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP32-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP32: liveins: $a0, $a1, $a2 ; MIPS32FP32-NEXT: liveins: $a0, $a1, $a2
; MIPS32FP32: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1 ; MIPS32FP32-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32FP32: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32FP32-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1
; MIPS32FP32: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32FP32-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32FP32: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP32-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32FP32: J %bb.2, implicit-def $at ; MIPS32FP32-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP32: bb.1.cond.true: ; MIPS32FP32-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: J %bb.3, implicit-def $at ; MIPS32FP32-NEXT: bb.1.cond.true:
; MIPS32FP32: bb.2.cond.false: ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: bb.3.cond.end: ; MIPS32FP32-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP32: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: $f0 = COPY [[PHI]] ; MIPS32FP32-NEXT: bb.2.cond.false:
; MIPS32FP32: RetRA implicit $f0 ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: bb.3.cond.end:
; MIPS32FP32-NEXT: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2
; MIPS32FP32-NEXT: $f0 = COPY [[PHI]]
; MIPS32FP32-NEXT: RetRA implicit $f0
;
; MIPS32FP64-LABEL: name: phi_float ; MIPS32FP64-LABEL: name: phi_float
; MIPS32FP64: bb.0.entry: ; MIPS32FP64: bb.0.entry:
; MIPS32FP64: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP64-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP64: liveins: $a0, $a1, $a2 ; MIPS32FP64-NEXT: liveins: $a0, $a1, $a2
; MIPS32FP64: [[COPY:%[0-9]+]]:gpr32 = COPY $a0 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1 ; MIPS32FP64-NEXT: [[COPY:%[0-9]+]]:gpr32 = COPY $a0
; MIPS32FP64: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2 ; MIPS32FP64-NEXT: [[COPY1:%[0-9]+]]:gpr32 = COPY $a1
; MIPS32FP64: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1 ; MIPS32FP64-NEXT: [[COPY2:%[0-9]+]]:gpr32 = COPY $a2
; MIPS32FP64: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP64-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[COPY]], 1
; MIPS32FP64: J %bb.2, implicit-def $at ; MIPS32FP64-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP64: bb.1.cond.true: ; MIPS32FP64-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: J %bb.3, implicit-def $at ; MIPS32FP64-NEXT: bb.1.cond.true:
; MIPS32FP64: bb.2.cond.false: ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: bb.3.cond.end: ; MIPS32FP64-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP64: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: $f0 = COPY [[PHI]] ; MIPS32FP64-NEXT: bb.2.cond.false:
; MIPS32FP64: RetRA implicit $f0 ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: bb.3.cond.end:
; MIPS32FP64-NEXT: [[PHI:%[0-9]+]]:gpr32 = PHI [[COPY1]], %bb.1, [[COPY2]], %bb.2
; MIPS32FP64-NEXT: $f0 = COPY [[PHI]]
; MIPS32FP64-NEXT: RetRA implicit $f0
bb.1.entry: bb.1.entry:
liveins: $a0, $a1, $a2 liveins: $a0, $a1, $a2
%3:gprb(s32) = COPY $a0 %3:gprb(s32) = COPY $a0
%1:gprb(s32) = COPY $a1 %1:gprb(s32) = COPY $a1
%2:gprb(s32) = COPY $a2 %2:gprb(s32) = COPY $a2
%6:gprb(s32) = G_CONSTANT i32 1 %6:gprb(s32) = G_CONSTANT i32 1
%7:gprb(s32) = COPY %3(s32) %7:gprb(s32) = COPY %3(s32)
Show All 18 Lines
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
fixedStack: fixedStack:
- { id: 0, offset: 16, size: 4, alignment: 8, isImmutable: true } - { id: 0, offset: 16, size: 4, alignment: 8, isImmutable: true }
body: | body: |
; MIPS32FP32-LABEL: name: phi_double ; MIPS32FP32-LABEL: name: phi_double
; MIPS32FP32: bb.0.entry: ; MIPS32FP32: bb.0.entry:
; MIPS32FP32: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP32-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP32: liveins: $d6, $d7 ; MIPS32FP32-NEXT: liveins: $d6, $d7
; MIPS32FP32: [[COPY:%[0-9]+]]:afgr64 = COPY $d6 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: [[COPY1:%[0-9]+]]:afgr64 = COPY $d7 ; MIPS32FP32-NEXT: [[COPY:%[0-9]+]]:afgr64 = COPY $d6
; MIPS32FP32: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0 ; MIPS32FP32-NEXT: [[COPY1:%[0-9]+]]:afgr64 = COPY $d7
; MIPS32FP32: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8) ; MIPS32FP32-NEXT: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0
; MIPS32FP32: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[LW]], 1 ; MIPS32FP32-NEXT: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8)
; MIPS32FP32: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP32-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[LW]], 1
; MIPS32FP32: J %bb.2, implicit-def $at ; MIPS32FP32-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP32: bb.1.cond.true: ; MIPS32FP32-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: J %bb.3, implicit-def $at ; MIPS32FP32-NEXT: bb.1.cond.true:
; MIPS32FP32: bb.2.cond.false: ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32: successors: %bb.3(0x80000000) ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: bb.3.cond.end: ; MIPS32FP32-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP32: [[PHI:%[0-9]+]]:afgr64 = PHI [[COPY]], %bb.1, [[COPY1]], %bb.2 ; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32: $d0 = COPY [[PHI]] ; MIPS32FP32-NEXT: bb.2.cond.false:
; MIPS32FP32: RetRA implicit $d0 ; MIPS32FP32-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: {{ $}}
; MIPS32FP32-NEXT: bb.3.cond.end:
; MIPS32FP32-NEXT: [[PHI:%[0-9]+]]:afgr64 = PHI [[COPY]], %bb.1, [[COPY1]], %bb.2
; MIPS32FP32-NEXT: $d0 = COPY [[PHI]]
; MIPS32FP32-NEXT: RetRA implicit $d0
;
; MIPS32FP64-LABEL: name: phi_double ; MIPS32FP64-LABEL: name: phi_double
; MIPS32FP64: bb.0.entry: ; MIPS32FP64: bb.0.entry:
; MIPS32FP64: successors: %bb.1(0x40000000), %bb.2(0x40000000) ; MIPS32FP64-NEXT: successors: %bb.1(0x40000000), %bb.2(0x40000000)
; MIPS32FP64: liveins: $d6, $d7 ; MIPS32FP64-NEXT: liveins: $d6, $d7
; MIPS32FP64: [[COPY:%[0-9]+]]:fgr64 = COPY $d6 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: [[COPY1:%[0-9]+]]:fgr64 = COPY $d7 ; MIPS32FP64-NEXT: [[COPY:%[0-9]+]]:fgr64 = COPY $d6
; MIPS32FP64: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0 ; MIPS32FP64-NEXT: [[COPY1:%[0-9]+]]:fgr64 = COPY $d7
; MIPS32FP64: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8) ; MIPS32FP64-NEXT: [[ADDiu:%[0-9]+]]:gpr32 = ADDiu %fixed-stack.0, 0
; MIPS32FP64: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[LW]], 1 ; MIPS32FP64-NEXT: [[LW:%[0-9]+]]:gpr32 = LW [[ADDiu]], 0 :: (load (s32) from %fixed-stack.0, align 8)
; MIPS32FP64: BNE [[ANDi]], $zero, %bb.1, implicit-def $at ; MIPS32FP64-NEXT: [[ANDi:%[0-9]+]]:gpr32 = ANDi [[LW]], 1
; MIPS32FP64: J %bb.2, implicit-def $at ; MIPS32FP64-NEXT: BNE [[ANDi]], $zero, %bb.1, implicit-def $at
; MIPS32FP64: bb.1.cond.true: ; MIPS32FP64-NEXT: J %bb.2, implicit-def dead $at
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: J %bb.3, implicit-def $at ; MIPS32FP64-NEXT: bb.1.cond.true:
; MIPS32FP64: bb.2.cond.false: ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64: successors: %bb.3(0x80000000) ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: bb.3.cond.end: ; MIPS32FP64-NEXT: J %bb.3, implicit-def dead $at
; MIPS32FP64: [[PHI:%[0-9]+]]:fgr64 = PHI [[COPY]], %bb.1, [[COPY1]], %bb.2 ; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64: $d0 = COPY [[PHI]] ; MIPS32FP64-NEXT: bb.2.cond.false:
; MIPS32FP64: RetRA implicit $d0 ; MIPS32FP64-NEXT: successors: %bb.3(0x80000000)
; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: {{ $}}
; MIPS32FP64-NEXT: bb.3.cond.end:
; MIPS32FP64-NEXT: [[PHI:%[0-9]+]]:fgr64 = PHI [[COPY]], %bb.1, [[COPY1]], %bb.2
; MIPS32FP64-NEXT: $d0 = COPY [[PHI]]
; MIPS32FP64-NEXT: RetRA implicit $d0
bb.1.entry: bb.1.entry:
liveins: $d6, $d7 liveins: $d6, $d7
%0:fprb(s64) = COPY $d6 %0:fprb(s64) = COPY $d6
%1:fprb(s64) = COPY $d7 %1:fprb(s64) = COPY $d7
%4:gprb(p0) = G_FRAME_INDEX %fixed-stack.0 %4:gprb(p0) = G_FRAME_INDEX %fixed-stack.0
%3:gprb(s32) = G_LOAD %4(p0) :: (load (s32) from %fixed-stack.0, align 8) %3:gprb(s32) = G_LOAD %4(p0) :: (load (s32) from %fixed-stack.0, align 8)
%7:gprb(s32) = G_CONSTANT i32 1 %7:gprb(s32) = G_CONSTANT i32 1
Show All 16 Lines

llvm/test/CodeGen/X86/GlobalISel/mul-scalar.ll

Show First 20 Lines • Show All 68 Lines • ▼ Show 20 Lines
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx
; X86-NEXT: movl {{[0-9]+}}(%esp), %esi ; X86-NEXT: movl {{[0-9]+}}(%esp), %esi
; X86-NEXT: imull %eax, %esi ; X86-NEXT: imull %eax, %esi
; X86-NEXT: movl %eax, %ecx ; X86-NEXT: movl %eax, %ecx
; X86-NEXT: imull %edx, %ecx ; X86-NEXT: imull %edx, %ecx
; X86-NEXT: movl {{[0-9]+}}(%esp), %edi ; X86-NEXT: movl {{[0-9]+}}(%esp), %edi
; X86-NEXT: imull %edx, %edi ; X86-NEXT: imull %edx, %edi
; X86-NEXT: mull %edx
; X86-NEXT: addl %edi, %esi ; X86-NEXT: addl %edi, %esi
; X86-NEXT: mull %edx
; X86-NEXT: addl %esi, %edx ; X86-NEXT: addl %esi, %edx
; X86-NEXT: movl %ecx, %eax ; X86-NEXT: movl %ecx, %eax
; X86-NEXT: popl %esi ; X86-NEXT: popl %esi
; X86-NEXT: popl %edi ; X86-NEXT: popl %edi
; X86-NEXT: retl ; X86-NEXT: retl
%ret = mul i64 %arg1, %arg2 %ret = mul i64 %arg1, %arg2
ret i64 %ret ret i64 %ret
} }
;TODO: instruction selection not supported yet ;TODO: instruction selection not supported yet
;define i128 @test_mul_i128(i128 %arg1, i128 %arg2) nounwind { ;define i128 @test_mul_i128(i128 %arg1, i128 %arg2) nounwind {
; %ret = mul i128 %arg1, %arg2 ; %ret = mul i128 %arg1, %arg2
; ret i128 %ret ; ret i128 %ret
;} ;}

llvm/test/CodeGen/X86/GlobalISel/select-add.mir

Show First 20 Lines • Show All 87 Lines • ▼ Show 20 Lines
regBankSelected: true regBankSelected: true
selected: false selected: false
registers: registers:
- { id: 0, class: gpr } - { id: 0, class: gpr }
- { id: 1, class: gpr } - { id: 1, class: gpr }
- { id: 2, class: gpr } - { id: 2, class: gpr }
# ALL: %0:gr16 = COPY $di # ALL: %0:gr16 = COPY $di
# ALL: %1:gr16 = COPY $si # ALL: %1:gr16 = COPY $si
# ALL: %2:gr16 = ADD16rr %0, %1, implicit-def $eflags # ALL: %2:gr16 = ADD16rr %0, %1, implicit-def dead $eflags
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
%0(s16) = COPY $di %0(s16) = COPY $di
%1(s16) = COPY $si %1(s16) = COPY $si
%2(s16) = G_ADD %0, %1 %2(s16) = G_ADD %0, %1
$ax = COPY %2(s16) $ax = COPY %2(s16)
RET 0, implicit $ax RET 0, implicit $ax
... ...
--- ---
name: test_add_i8 name: test_add_i8
# ALL-LABEL: name: test_add_i8 # ALL-LABEL: name: test_add_i8
alignment: 16 alignment: 16
legalized: true legalized: true
regBankSelected: true regBankSelected: true
selected: false selected: false
registers: registers:
- { id: 0, class: gpr } - { id: 0, class: gpr }
- { id: 1, class: gpr } - { id: 1, class: gpr }
- { id: 2, class: gpr } - { id: 2, class: gpr }
# ALL: %0:gr8 = COPY $dil # ALL: %0:gr8 = COPY $dil
# ALL: %1:gr8 = COPY $sil # ALL: %1:gr8 = COPY $sil
# ALL: %2:gr8 = ADD8rr %0, %1, implicit-def $eflags # ALL: %2:gr8 = ADD8rr %0, %1, implicit-def dead $eflags
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
%0(s8) = COPY $dil %0(s8) = COPY $dil
%1(s8) = COPY $sil %1(s8) = COPY $sil
%2(s8) = G_ADD %0, %1 %2(s8) = G_ADD %0, %1
$al = COPY %2(s8) $al = COPY %2(s8)
▲ Show 20 Lines • Show All 69 Lines • Show Last 20 Lines

llvm/test/CodeGen/X86/GlobalISel/select-and-scalar.mir

Show All 35 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_and_i8 ; ALL-LABEL: name: test_and_i8
; ALL: [[COPY:%[0-9]+]]:gr8 = COPY $dil ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr8 = COPY $sil ; ALL-NEXT: {{ $}}
; ALL: [[AND8rr:%[0-9]+]]:gr8 = AND8rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr8 = COPY $dil
; ALL: $al = COPY [[AND8rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr8 = COPY $sil
; ALL: RET 0, implicit $al ; ALL-NEXT: [[AND8rr:%[0-9]+]]:gr8 = AND8rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $al = COPY [[AND8rr]]
; ALL-NEXT: RET 0, implicit $al
%0(s8) = COPY $dil %0(s8) = COPY $dil
%1(s8) = COPY $sil %1(s8) = COPY $sil
%2(s8) = G_AND %0, %1 %2(s8) = G_AND %0, %1
$al = COPY %2(s8) $al = COPY %2(s8)
RET 0, implicit $al RET 0, implicit $al
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_and_i16 ; ALL-LABEL: name: test_and_i16
; ALL: [[COPY:%[0-9]+]]:gr16 = COPY $di ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr16 = COPY $si ; ALL-NEXT: {{ $}}
; ALL: [[AND16rr:%[0-9]+]]:gr16 = AND16rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr16 = COPY $di
; ALL: $ax = COPY [[AND16rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr16 = COPY $si
; ALL: RET 0, implicit $ax ; ALL-NEXT: [[AND16rr:%[0-9]+]]:gr16 = AND16rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $ax = COPY [[AND16rr]]
; ALL-NEXT: RET 0, implicit $ax
%0(s16) = COPY $di %0(s16) = COPY $di
%1(s16) = COPY $si %1(s16) = COPY $si
%2(s16) = G_AND %0, %1 %2(s16) = G_AND %0, %1
$ax = COPY %2(s16) $ax = COPY %2(s16)
RET 0, implicit $ax RET 0, implicit $ax
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_and_i32 ; ALL-LABEL: name: test_and_i32
; ALL: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr32 = COPY $esi ; ALL-NEXT: {{ $}}
; ALL: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; ALL: $eax = COPY [[AND32rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr32 = COPY $esi
; ALL: RET 0, implicit $eax ; ALL-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $eax = COPY [[AND32rr]]
; ALL-NEXT: RET 0, implicit $eax
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = COPY $esi %1(s32) = COPY $esi
%2(s32) = G_AND %0, %1 %2(s32) = G_AND %0, %1
$eax = COPY %2(s32) $eax = COPY %2(s32)
RET 0, implicit $eax RET 0, implicit $eax
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $rdi, $rsi liveins: $rdi, $rsi
; ALL-LABEL: name: test_and_i64 ; ALL-LABEL: name: test_and_i64
; ALL: [[COPY:%[0-9]+]]:gr64 = COPY $rdi ; ALL: liveins: $rdi, $rsi
; ALL: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi ; ALL-NEXT: {{ $}}
; ALL: [[AND64rr:%[0-9]+]]:gr64 = AND64rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr64 = COPY $rdi
; ALL: $rax = COPY [[AND64rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi
; ALL: RET 0, implicit $rax ; ALL-NEXT: [[AND64rr:%[0-9]+]]:gr64 = AND64rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $rax = COPY [[AND64rr]]
; ALL-NEXT: RET 0, implicit $rax
%0(s64) = COPY $rdi %0(s64) = COPY $rdi
%1(s64) = COPY $rsi %1(s64) = COPY $rsi
%2(s64) = G_AND %0, %1 %2(s64) = G_AND %0, %1
$rax = COPY %2(s64) $rax = COPY %2(s64)
RET 0, implicit $rax RET 0, implicit $rax
... ...

llvm/test/CodeGen/X86/GlobalISel/select-blsi.mir

Show All 22 Linesbody: |
bb.1: bb.1:
liveins: $edi liveins: $edi
; CHECK-LABEL: name: test_blsi32rr ; CHECK-LABEL: name: test_blsi32rr
; CHECK: liveins: $edi ; CHECK: liveins: $edi
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; CHECK-NEXT: [[MOV32r0_:%[0-9]+]]:gr32 = MOV32r0 implicit-def $eflags ; CHECK-NEXT: [[MOV32r0_:%[0-9]+]]:gr32 = MOV32r0 implicit-def $eflags
; CHECK-NEXT: [[SUB32rr:%[0-9]+]]:gr32 = SUB32rr [[MOV32r0_]], [[COPY]], implicit-def $eflags ; CHECK-NEXT: [[SUB32rr:%[0-9]+]]:gr32 = SUB32rr [[MOV32r0_]], [[COPY]], implicit-def dead $eflags
; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[SUB32rr]], [[COPY]], implicit-def $eflags ; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[SUB32rr]], [[COPY]], implicit-def dead $eflags
; CHECK-NEXT: $edi = COPY [[AND32rr]] ; CHECK-NEXT: $edi = COPY [[AND32rr]]
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = G_CONSTANT i32 0 %1(s32) = G_CONSTANT i32 0
%2(s32) = G_SUB %1, %0 %2(s32) = G_SUB %1, %0
%3(s32) = G_AND %2, %0 %3(s32) = G_AND %2, %0
$edi = COPY %3 $edi = COPY %3
... ...
Show All 13 Lines bb.1:
liveins: $edi liveins: $edi
; CHECK-LABEL: name: test_blsi32rr_nomatch ; CHECK-LABEL: name: test_blsi32rr_nomatch
; CHECK: liveins: $edi ; CHECK: liveins: $edi
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; CHECK-NEXT: [[MOV32r0_:%[0-9]+]]:gr32 = MOV32r0 implicit-def $eflags ; CHECK-NEXT: [[MOV32r0_:%[0-9]+]]:gr32 = MOV32r0 implicit-def $eflags
; CHECK-NEXT: [[SUB32ri:%[0-9]+]]:gr32 = SUB32ri [[MOV32r0_]], 0, implicit-def $eflags ; CHECK-NEXT: [[SUB32ri:%[0-9]+]]:gr32 = SUB32ri [[MOV32r0_]], 0, implicit-def $eflags
; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[SUB32ri]], [[COPY]], implicit-def $eflags ; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[SUB32ri]], [[COPY]], implicit-def dead $eflags
; CHECK-NEXT: $edi = COPY [[AND32rr]] ; CHECK-NEXT: $edi = COPY [[AND32rr]]
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = G_CONSTANT i32 0 %1(s32) = G_CONSTANT i32 0
%2(s32) = G_SUB %1, %1 %2(s32) = G_SUB %1, %1
%3(s32) = G_AND %2, %0 %3(s32) = G_AND %2, %0
$edi = COPY %3 $edi = COPY %3
... ...

llvm/test/CodeGen/X86/GlobalISel/select-blsr.mir

Show All 19 Linesbody: |
bb.1: bb.1:
liveins: $edi liveins: $edi
; CHECK-LABEL: name: test_blsr32rr ; CHECK-LABEL: name: test_blsr32rr
; CHECK: liveins: $edi ; CHECK: liveins: $edi
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; CHECK-NEXT: [[DEC32r:%[0-9]+]]:gr32 = DEC32r [[COPY]], implicit-def $eflags ; CHECK-NEXT: [[DEC32r:%[0-9]+]]:gr32 = DEC32r [[COPY]], implicit-def $eflags
; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[DEC32r]], [[COPY]], implicit-def $eflags ; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[DEC32r]], [[COPY]], implicit-def dead $eflags
; CHECK-NEXT: $edi = COPY [[AND32rr]] ; CHECK-NEXT: $edi = COPY [[AND32rr]]
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = G_CONSTANT i32 -1 %1(s32) = G_CONSTANT i32 -1
%2(s32) = G_ADD %0, %1 %2(s32) = G_ADD %0, %1
%3(s32) = G_AND %2, %0 %3(s32) = G_AND %2, %0
$edi = COPY %3 $edi = COPY %3
... ...
Show All 13 Lines bb.1:
liveins: $edi liveins: $edi
; CHECK-LABEL: name: test_blsr32rr_nomatch ; CHECK-LABEL: name: test_blsr32rr_nomatch
; CHECK: liveins: $edi ; CHECK: liveins: $edi
; CHECK-NEXT: {{ $}} ; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; CHECK-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; CHECK-NEXT: [[MOV32ri:%[0-9]+]]:gr32 = MOV32ri -1 ; CHECK-NEXT: [[MOV32ri:%[0-9]+]]:gr32 = MOV32ri -1
; CHECK-NEXT: [[DEC32r:%[0-9]+]]:gr32 = DEC32r [[MOV32ri]], implicit-def $eflags ; CHECK-NEXT: [[DEC32r:%[0-9]+]]:gr32 = DEC32r [[MOV32ri]], implicit-def $eflags
; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[DEC32r]], [[COPY]], implicit-def $eflags ; CHECK-NEXT: [[AND32rr:%[0-9]+]]:gr32 = AND32rr [[DEC32r]], [[COPY]], implicit-def dead $eflags
; CHECK-NEXT: $edi = COPY [[AND32rr]] ; CHECK-NEXT: $edi = COPY [[AND32rr]]
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = G_CONSTANT i32 -1 %1(s32) = G_CONSTANT i32 -1
%2(s32) = G_ADD %1, %1 %2(s32) = G_ADD %1, %1
%3(s32) = G_AND %2, %0 %3(s32) = G_AND %2, %0
$edi = COPY %3 $edi = COPY %3
... ...

llvm/test/CodeGen/X86/GlobalISel/select-mul-scalar.mir

Show All 26 Linesregisters:
- { id: 0, class: gpr } - { id: 0, class: gpr }
- { id: 1, class: gpr } - { id: 1, class: gpr }
- { id: 2, class: gpr } - { id: 2, class: gpr }
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_mul_i16 ; ALL-LABEL: name: test_mul_i16
; ALL: [[COPY:%[0-9]+]]:gr16 = COPY $di ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr16 = COPY $si ; ALL-NEXT: {{ $}}
; ALL: [[IMUL16rr:%[0-9]+]]:gr16 = IMUL16rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr16 = COPY $di
; ALL: $ax = COPY [[IMUL16rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr16 = COPY $si
; ALL: RET 0, implicit $ax ; ALL-NEXT: [[IMUL16rr:%[0-9]+]]:gr16 = IMUL16rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $ax = COPY [[IMUL16rr]]
; ALL-NEXT: RET 0, implicit $ax
%0(s16) = COPY $di %0(s16) = COPY $di
%1(s16) = COPY $si %1(s16) = COPY $si
%2(s16) = G_MUL %0, %1 %2(s16) = G_MUL %0, %1
$ax = COPY %2(s16) $ax = COPY %2(s16)
RET 0, implicit $ax RET 0, implicit $ax
... ...
--- ---
name: test_mul_i32 name: test_mul_i32
alignment: 16 alignment: 16
legalized: true legalized: true
regBankSelected: true regBankSelected: true
registers: registers:
- { id: 0, class: gpr } - { id: 0, class: gpr }
- { id: 1, class: gpr } - { id: 1, class: gpr }
- { id: 2, class: gpr } - { id: 2, class: gpr }
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_mul_i32 ; ALL-LABEL: name: test_mul_i32
; ALL: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr32 = COPY $esi ; ALL-NEXT: {{ $}}
; ALL: [[IMUL32rr:%[0-9]+]]:gr32 = IMUL32rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; ALL: $eax = COPY [[IMUL32rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr32 = COPY $esi
; ALL: RET 0, implicit $eax ; ALL-NEXT: [[IMUL32rr:%[0-9]+]]:gr32 = IMUL32rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $eax = COPY [[IMUL32rr]]
; ALL-NEXT: RET 0, implicit $eax
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = COPY $esi %1(s32) = COPY $esi
%2(s32) = G_MUL %0, %1 %2(s32) = G_MUL %0, %1
$eax = COPY %2(s32) $eax = COPY %2(s32)
RET 0, implicit $eax RET 0, implicit $eax
... ...
--- ---
name: test_mul_i64 name: test_mul_i64
alignment: 16 alignment: 16
legalized: true legalized: true
regBankSelected: true regBankSelected: true
registers: registers:
- { id: 0, class: gpr } - { id: 0, class: gpr }
- { id: 1, class: gpr } - { id: 1, class: gpr }
- { id: 2, class: gpr } - { id: 2, class: gpr }
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $rdi, $rsi liveins: $rdi, $rsi
; ALL-LABEL: name: test_mul_i64 ; ALL-LABEL: name: test_mul_i64
; ALL: [[COPY:%[0-9]+]]:gr64 = COPY $rdi ; ALL: liveins: $rdi, $rsi
; ALL: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi ; ALL-NEXT: {{ $}}
; ALL: [[IMUL64rr:%[0-9]+]]:gr64 = IMUL64rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr64 = COPY $rdi
; ALL: $rax = COPY [[IMUL64rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi
; ALL: RET 0, implicit $rax ; ALL-NEXT: [[IMUL64rr:%[0-9]+]]:gr64 = IMUL64rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $rax = COPY [[IMUL64rr]]
; ALL-NEXT: RET 0, implicit $rax
%0(s64) = COPY $rdi %0(s64) = COPY $rdi
%1(s64) = COPY $rsi %1(s64) = COPY $rsi
%2(s64) = G_MUL %0, %1 %2(s64) = G_MUL %0, %1
$rax = COPY %2(s64) $rax = COPY %2(s64)
RET 0, implicit $rax RET 0, implicit $rax
... ...

llvm/test/CodeGen/X86/GlobalISel/select-or-scalar.mir

Show All 35 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_or_i8 ; ALL-LABEL: name: test_or_i8
; ALL: [[COPY:%[0-9]+]]:gr8 = COPY $dil ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr8 = COPY $sil ; ALL-NEXT: {{ $}}
; ALL: [[OR8rr:%[0-9]+]]:gr8 = OR8rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr8 = COPY $dil
; ALL: $al = COPY [[OR8rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr8 = COPY $sil
; ALL: RET 0, implicit $al ; ALL-NEXT: [[OR8rr:%[0-9]+]]:gr8 = OR8rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $al = COPY [[OR8rr]]
; ALL-NEXT: RET 0, implicit $al
%0(s8) = COPY $dil %0(s8) = COPY $dil
%1(s8) = COPY $sil %1(s8) = COPY $sil
%2(s8) = G_OR %0, %1 %2(s8) = G_OR %0, %1
$al = COPY %2(s8) $al = COPY %2(s8)
RET 0, implicit $al RET 0, implicit $al
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_or_i16 ; ALL-LABEL: name: test_or_i16
; ALL: [[COPY:%[0-9]+]]:gr16 = COPY $di ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr16 = COPY $si ; ALL-NEXT: {{ $}}
; ALL: [[OR16rr:%[0-9]+]]:gr16 = OR16rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr16 = COPY $di
; ALL: $ax = COPY [[OR16rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr16 = COPY $si
; ALL: RET 0, implicit $ax ; ALL-NEXT: [[OR16rr:%[0-9]+]]:gr16 = OR16rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $ax = COPY [[OR16rr]]
; ALL-NEXT: RET 0, implicit $ax
%0(s16) = COPY $di %0(s16) = COPY $di
%1(s16) = COPY $si %1(s16) = COPY $si
%2(s16) = G_OR %0, %1 %2(s16) = G_OR %0, %1
$ax = COPY %2(s16) $ax = COPY %2(s16)
RET 0, implicit $ax RET 0, implicit $ax
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_or_i32 ; ALL-LABEL: name: test_or_i32
; ALL: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr32 = COPY $esi ; ALL-NEXT: {{ $}}
; ALL: [[OR32rr:%[0-9]+]]:gr32 = OR32rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; ALL: $eax = COPY [[OR32rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr32 = COPY $esi
; ALL: RET 0, implicit $eax ; ALL-NEXT: [[OR32rr:%[0-9]+]]:gr32 = OR32rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $eax = COPY [[OR32rr]]
; ALL-NEXT: RET 0, implicit $eax
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = COPY $esi %1(s32) = COPY $esi
%2(s32) = G_OR %0, %1 %2(s32) = G_OR %0, %1
$eax = COPY %2(s32) $eax = COPY %2(s32)
RET 0, implicit $eax RET 0, implicit $eax
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $rdi, $rsi liveins: $rdi, $rsi
; ALL-LABEL: name: test_or_i64 ; ALL-LABEL: name: test_or_i64
; ALL: [[COPY:%[0-9]+]]:gr64 = COPY $rdi ; ALL: liveins: $rdi, $rsi
; ALL: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi ; ALL-NEXT: {{ $}}
; ALL: [[OR64rr:%[0-9]+]]:gr64 = OR64rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr64 = COPY $rdi
; ALL: $rax = COPY [[OR64rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi
; ALL: RET 0, implicit $rax ; ALL-NEXT: [[OR64rr:%[0-9]+]]:gr64 = OR64rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $rax = COPY [[OR64rr]]
; ALL-NEXT: RET 0, implicit $rax
%0(s64) = COPY $rdi %0(s64) = COPY $rdi
%1(s64) = COPY $rsi %1(s64) = COPY $rsi
%2(s64) = G_OR %0, %1 %2(s64) = G_OR %0, %1
$rax = COPY %2(s64) $rax = COPY %2(s64)
RET 0, implicit $rax RET 0, implicit $rax
... ...

llvm/test/CodeGen/X86/GlobalISel/select-undef.mir

Show All 24 Lines
liveins: liveins:
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
; ALL-LABEL: name: test ; ALL-LABEL: name: test
; ALL: [[DEF:%[0-9]+]]:gr8 = IMPLICIT_DEF ; ALL: [[DEF:%[0-9]+]]:gr8 = IMPLICIT_DEF
; ALL: $al = COPY [[DEF]] ; ALL-NEXT: $al = COPY [[DEF]]
; ALL: RET 0, implicit $al ; ALL-NEXT: RET 0, implicit $al
%0(s8) = G_IMPLICIT_DEF %0(s8) = G_IMPLICIT_DEF
$al = COPY %0(s8) $al = COPY %0(s8)
RET 0, implicit $al RET 0, implicit $al
... ...
--- ---
name: test2 name: test2
alignment: 16 alignment: 16
legalized: true legalized: true
regBankSelected: true regBankSelected: true
registers: registers:
- { id: 0, class: gpr, preferred-register: '' } - { id: 0, class: gpr, preferred-register: '' }
- { id: 1, class: gpr, preferred-register: '' } - { id: 1, class: gpr, preferred-register: '' }
- { id: 2, class: gpr, preferred-register: '' } - { id: 2, class: gpr, preferred-register: '' }
liveins: liveins:
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi liveins: $edi
; ALL-LABEL: name: test2 ; ALL-LABEL: name: test2
; ALL: [[COPY:%[0-9]+]]:gr8 = COPY $dil ; ALL: liveins: $edi
; ALL: [[DEF:%[0-9]+]]:gr8 = IMPLICIT_DEF ; ALL-NEXT: {{ $}}
; ALL: [[ADD8rr:%[0-9]+]]:gr8 = ADD8rr [[COPY]], [[DEF]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr8 = COPY $dil
; ALL: $al = COPY [[ADD8rr]] ; ALL-NEXT: [[DEF:%[0-9]+]]:gr8 = IMPLICIT_DEF
; ALL: RET 0, implicit $al ; ALL-NEXT: [[ADD8rr:%[0-9]+]]:gr8 = ADD8rr [[COPY]], [[DEF]], implicit-def dead $eflags
; ALL-NEXT: $al = COPY [[ADD8rr]]
; ALL-NEXT: RET 0, implicit $al
%0(s8) = COPY $dil %0(s8) = COPY $dil
%1(s8) = G_IMPLICIT_DEF %1(s8) = G_IMPLICIT_DEF
%2(s8) = G_ADD %0, %1 %2(s8) = G_ADD %0, %1
$al = COPY %2(s8) $al = COPY %2(s8)
RET 0, implicit $al RET 0, implicit $al
... ...
--- ---
name: test3 name: test3
alignment: 16 alignment: 16
legalized: true legalized: true
regBankSelected: true regBankSelected: true
tracksRegLiveness: true tracksRegLiveness: true
registers: registers:
- { id: 1, class: vecr } - { id: 1, class: vecr }
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
; ALL-LABEL: name: test3 ; ALL-LABEL: name: test3
; ALL: [[DEF:%[0-9]+]]:vr128 = IMPLICIT_DEF ; ALL: [[DEF:%[0-9]+]]:vr128 = IMPLICIT_DEF
; ALL: $xmm0 = COPY [[DEF]] ; ALL-NEXT: $xmm0 = COPY [[DEF]]
; ALL: RET 0, implicit $xmm0 ; ALL-NEXT: RET 0, implicit $xmm0
%1:vecr(s128) = G_IMPLICIT_DEF %1:vecr(s128) = G_IMPLICIT_DEF
$xmm0 = COPY %1:vecr(s128) $xmm0 = COPY %1:vecr(s128)
RET 0, implicit $xmm0 RET 0, implicit $xmm0
... ...

llvm/test/CodeGen/X86/GlobalISel/select-xor-scalar.mir

Show All 35 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_xor_i8 ; ALL-LABEL: name: test_xor_i8
; ALL: [[COPY:%[0-9]+]]:gr8 = COPY $dil ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr8 = COPY $sil ; ALL-NEXT: {{ $}}
; ALL: [[XOR8rr:%[0-9]+]]:gr8 = XOR8rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr8 = COPY $dil
; ALL: $al = COPY [[XOR8rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr8 = COPY $sil
; ALL: RET 0, implicit $al ; ALL-NEXT: [[XOR8rr:%[0-9]+]]:gr8 = XOR8rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $al = COPY [[XOR8rr]]
; ALL-NEXT: RET 0, implicit $al
%0(s8) = COPY $dil %0(s8) = COPY $dil
%1(s8) = COPY $sil %1(s8) = COPY $sil
%2(s8) = G_XOR %0, %1 %2(s8) = G_XOR %0, %1
$al = COPY %2(s8) $al = COPY %2(s8)
RET 0, implicit $al RET 0, implicit $al
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_xor_i16 ; ALL-LABEL: name: test_xor_i16
; ALL: [[COPY:%[0-9]+]]:gr16 = COPY $di ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr16 = COPY $si ; ALL-NEXT: {{ $}}
; ALL: [[XOR16rr:%[0-9]+]]:gr16 = XOR16rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr16 = COPY $di
; ALL: $ax = COPY [[XOR16rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr16 = COPY $si
; ALL: RET 0, implicit $ax ; ALL-NEXT: [[XOR16rr:%[0-9]+]]:gr16 = XOR16rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $ax = COPY [[XOR16rr]]
; ALL-NEXT: RET 0, implicit $ax
%0(s16) = COPY $di %0(s16) = COPY $di
%1(s16) = COPY $si %1(s16) = COPY $si
%2(s16) = G_XOR %0, %1 %2(s16) = G_XOR %0, %1
$ax = COPY %2(s16) $ax = COPY %2(s16)
RET 0, implicit $ax RET 0, implicit $ax
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $edi, $esi liveins: $edi, $esi
; ALL-LABEL: name: test_xor_i32 ; ALL-LABEL: name: test_xor_i32
; ALL: [[COPY:%[0-9]+]]:gr32 = COPY $edi ; ALL: liveins: $edi, $esi
; ALL: [[COPY1:%[0-9]+]]:gr32 = COPY $esi ; ALL-NEXT: {{ $}}
; ALL: [[XOR32rr:%[0-9]+]]:gr32 = XOR32rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr32 = COPY $edi
; ALL: $eax = COPY [[XOR32rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr32 = COPY $esi
; ALL: RET 0, implicit $eax ; ALL-NEXT: [[XOR32rr:%[0-9]+]]:gr32 = XOR32rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $eax = COPY [[XOR32rr]]
; ALL-NEXT: RET 0, implicit $eax
%0(s32) = COPY $edi %0(s32) = COPY $edi
%1(s32) = COPY $esi %1(s32) = COPY $esi
%2(s32) = G_XOR %0, %1 %2(s32) = G_XOR %0, %1
$eax = COPY %2(s32) $eax = COPY %2(s32)
RET 0, implicit $eax RET 0, implicit $eax
... ...
--- ---
Show All 9 Lines
fixedStack: fixedStack:
stack: stack:
constants: constants:
body: | body: |
bb.1 (%ir-block.0): bb.1 (%ir-block.0):
liveins: $rdi, $rsi liveins: $rdi, $rsi
; ALL-LABEL: name: test_xor_i64 ; ALL-LABEL: name: test_xor_i64
; ALL: [[COPY:%[0-9]+]]:gr64 = COPY $rdi ; ALL: liveins: $rdi, $rsi
; ALL: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi ; ALL-NEXT: {{ $}}
; ALL: [[XOR64rr:%[0-9]+]]:gr64 = XOR64rr [[COPY]], [[COPY1]], implicit-def $eflags ; ALL-NEXT: [[COPY:%[0-9]+]]:gr64 = COPY $rdi
; ALL: $rax = COPY [[XOR64rr]] ; ALL-NEXT: [[COPY1:%[0-9]+]]:gr64 = COPY $rsi
; ALL: RET 0, implicit $rax ; ALL-NEXT: [[XOR64rr:%[0-9]+]]:gr64 = XOR64rr [[COPY]], [[COPY1]], implicit-def dead $eflags
; ALL-NEXT: $rax = COPY [[XOR64rr]]
; ALL-NEXT: RET 0, implicit $rax
%0(s64) = COPY $rdi %0(s64) = COPY $rdi
%1(s64) = COPY $rsi %1(s64) = COPY $rsi
%2(s64) = G_XOR %0, %1 %2(s64) = G_XOR %0, %1
$rax = COPY %2(s64) $rax = COPY %2(s64)
RET 0, implicit $rax RET 0, implicit $rax
... ...

llvm/utils/TableGen/GlobalISelEmitter.cpp

Show First 20 Lines • Show All 306 Lines • ▼ Show 20 Linespublic:
void emitMIPredicateFns(raw_ostream &OS) override; void emitMIPredicateFns(raw_ostream &OS) override;
void emitI64ImmPredicateFns(raw_ostream &OS) override; void emitI64ImmPredicateFns(raw_ostream &OS) override;
void emitAPFloatImmPredicateFns(raw_ostream &OS) override; void emitAPFloatImmPredicateFns(raw_ostream &OS) override;
void emitAPIntImmPredicateFns(raw_ostream &OS) override; void emitAPIntImmPredicateFns(raw_ostream &OS) override;
void emitTestSimplePredicate(raw_ostream &OS) override; void emitTestSimplePredicate(raw_ostream &OS) override;
void emitRunCustomAction(raw_ostream &OS) override; void emitRunCustomAction(raw_ostream &OS) override;
void postProcessRule(RuleMatcher &M);
const CodeGenTarget &getTarget() const override { return Target; } const CodeGenTarget &getTarget() const override { return Target; }
StringRef getClassName() const override { return ClassName; } StringRef getClassName() const override { return ClassName; }
void run(raw_ostream &OS); void run(raw_ostream &OS);
private: private:
std::string ClassName; std::string ClassName;
▲ Show 20 Lines • Show All 2,042 Lines • ▼ Show 20 Lines
void GlobalISelEmitter::emitRunCustomAction(raw_ostream &OS) { void GlobalISelEmitter::emitRunCustomAction(raw_ostream &OS) {
OS << "void " << getClassName() OS << "void " << getClassName()
<< "::runCustomAction(unsigned, const MatcherState&) const {\n" << "::runCustomAction(unsigned, const MatcherState&) const {\n"
<< " llvm_unreachable(\"" + getClassName() + << " llvm_unreachable(\"" + getClassName() +
" does not support custom C++ actions!\");\n" " does not support custom C++ actions!\");\n"
<< "}\n"; << "}\n";
} }
void GlobalISelEmitter::postProcessRule(RuleMatcher &M) {
SmallPtrSet<Record *, 16> UsedRegs;
// TODO: deal with subregs?
for (auto &A : M.actions()) {
auto *MI = dyn_cast<BuildMIAction>(A.get());
if (!MI)
continue;
for (auto *Use : MI->getCGI()->ImplicitUses)
UsedRegs.insert(Use);
}
for (auto &A : M.actions()) {
auto *MI = dyn_cast<BuildMIAction>(A.get());
if (!MI)
continue;
for (auto *Def : MI->getCGI()->ImplicitDefs) {
if (!UsedRegs.contains(Def))
MI->setDeadImplicitDef(Def);
}
}
}
void GlobalISelEmitter::run(raw_ostream &OS) { void GlobalISelEmitter::run(raw_ostream &OS) {
if (!UseCoverageFile.empty()) { if (!UseCoverageFile.empty()) {
RuleCoverage = CodeGenCoverage(); RuleCoverage = CodeGenCoverage();
auto RuleCoverageBufOrErr = MemoryBuffer::getFile(UseCoverageFile); auto RuleCoverageBufOrErr = MemoryBuffer::getFile(UseCoverageFile);
if (!RuleCoverageBufOrErr) { if (!RuleCoverageBufOrErr) {
PrintWarning(SMLoc(), "Missing rule coverage data"); PrintWarning(SMLoc(), "Missing rule coverage data");
RuleCoverage = std::nullopt; RuleCoverage = std::nullopt;
} else { } else {
▲ Show 20 Lines • Show All 41 Lines • ▼ Show 20 Lines for (const PatternToMatch &Pat : CGP.ptms()) {
if (RuleCoverage) { if (RuleCoverage) {
if (RuleCoverage->isCovered(MatcherOrErr->getRuleID())) if (RuleCoverage->isCovered(MatcherOrErr->getRuleID()))
++NumPatternsTested; ++NumPatternsTested;
else else
PrintWarning(Pat.getSrcRecord()->getLoc(), PrintWarning(Pat.getSrcRecord()->getLoc(),
"Pattern is not covered by a test"); "Pattern is not covered by a test");
} }
Rules.push_back(std::move(MatcherOrErr.get())); Rules.push_back(std::move(MatcherOrErr.get()));
postProcessRule(Rules.back());
} }
// Comparison function to order records by name. // Comparison function to order records by name.
auto orderByName = [](const Record *A, const Record *B) { auto orderByName = [](const Record *A, const Record *B) {
return A->getName() < B->getName(); return A->getName() < B->getName();
}; };
std::vector<Record *> ComplexPredicates = std::vector<Record *> ComplexPredicates =
▲ Show 20 Lines • Show All 67 Lines • Show Last 20 Lines

llvm/utils/TableGen/GlobalISelMatchTable.h

Show First 20 Lines • Show All 2,045 Lines • ▼ Show 20 Lines
/// An action taken when all Matcher predicates succeeded for a parent rule. /// An action taken when all Matcher predicates succeeded for a parent rule.
/// ///
/// Typical actions include: /// Typical actions include:
/// * Changing the opcode of an instruction. /// * Changing the opcode of an instruction.
/// * Adding an operand to an instruction. /// * Adding an operand to an instruction.
class MatchAction { class MatchAction {
public: public:
enum ActionKind {
AK_DebugComment,
AK_CustomCXX,
AK_BuildMI,
AK_ConstraintOpsToDef,
AK_ConstraintOpsToRC,
AK_MakeTempReg,
};
MatchAction(ActionKind K) : Kind(K) {}
ActionKind getKind() const { return Kind; }
virtual ~MatchAction() {} virtual ~MatchAction() {}
/// Emit the MatchTable opcodes to implement the action. /// Emit the MatchTable opcodes to implement the action.
virtual void emitActionOpcodes(MatchTable &Table, virtual void emitActionOpcodes(MatchTable &Table,
RuleMatcher &Rule) const = 0; RuleMatcher &Rule) const = 0;
private:
ActionKind Kind;
}; };
/// Generates a comment describing the matched rule being acted upon. /// Generates a comment describing the matched rule being acted upon.
class DebugCommentAction : public MatchAction { class DebugCommentAction : public MatchAction {
private: private:
std::string S; std::string S;
public: public:
DebugCommentAction(StringRef S) : S(std::string(S)) {} DebugCommentAction(StringRef S)
: MatchAction(AK_DebugComment), S(std::string(S)) {}
static bool classof(const MatchAction *A) {
return A->getKind() == AK_DebugComment;
}
void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override { void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override {
Table << MatchTable::Comment(S) << MatchTable::LineBreak; Table << MatchTable::Comment(S) << MatchTable::LineBreak;
} }
}; };
class CustomCXXAction : public MatchAction { class CustomCXXAction : public MatchAction {
std::string FnEnumName; std::string FnEnumName;
public: public:
CustomCXXAction(StringRef FnEnumName) : FnEnumName(FnEnumName.str()) {} CustomCXXAction(StringRef FnEnumName)
: MatchAction(AK_CustomCXX), FnEnumName(FnEnumName.str()) {}
static bool classof(const MatchAction *A) {
return A->getKind() == AK_CustomCXX;
}
void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override; void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override;
}; };
/// Generates code to build an instruction or mutate an existing instruction /// Generates code to build an instruction or mutate an existing instruction
/// into the desired instruction when this is possible. /// into the desired instruction when this is possible.
class BuildMIAction : public MatchAction { class BuildMIAction : public MatchAction {
private: private:
unsigned InsnID; unsigned InsnID;
const CodeGenInstruction *I; const CodeGenInstruction *I;
InstructionMatcher *Matched; InstructionMatcher *Matched;
std::vector<std::unique_ptr<OperandRenderer>> OperandRenderers; std::vector<std::unique_ptr<OperandRenderer>> OperandRenderers;
SmallPtrSet<Record *, 4> DeadImplicitDefs;
/// True if the instruction can be built solely by mutating the opcode. /// True if the instruction can be built solely by mutating the opcode.
bool canMutate(RuleMatcher &Rule, const InstructionMatcher *Insn) const; bool canMutate(RuleMatcher &Rule, const InstructionMatcher *Insn) const;
public: public:
BuildMIAction(unsigned InsnID, const CodeGenInstruction *I) BuildMIAction(unsigned InsnID, const CodeGenInstruction *I)
: InsnID(InsnID), I(I), Matched(nullptr) {} : MatchAction(AK_BuildMI), InsnID(InsnID), I(I), Matched(nullptr) {}
static bool classof(const MatchAction *A) {
return A->getKind() == AK_BuildMI;
}
unsigned getInsnID() const { return InsnID; } unsigned getInsnID() const { return InsnID; }
const CodeGenInstruction *getCGI() const { return I; } const CodeGenInstruction *getCGI() const { return I; }
void chooseInsnToMutate(RuleMatcher &Rule); void chooseInsnToMutate(RuleMatcher &Rule);
void setDeadImplicitDef(Record *R) { DeadImplicitDefs.insert(R); }
template <class Kind, class... Args> Kind &addRenderer(Args &&...args) { template <class Kind, class... Args> Kind &addRenderer(Args &&...args) {
OperandRenderers.emplace_back( OperandRenderers.emplace_back(
std::make_unique<Kind>(InsnID, std::forward<Args>(args)...)); std::make_unique<Kind>(InsnID, std::forward<Args>(args)...));
return *static_cast<Kind *>(OperandRenderers.back().get()); return *static_cast<Kind *>(OperandRenderers.back().get());
} }
void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override; void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override;
}; };
/// Generates code to constrain the operands of an output instruction to the /// Generates code to constrain the operands of an output instruction to the
/// register classes specified by the definition of that instruction. /// register classes specified by the definition of that instruction.
class ConstrainOperandsToDefinitionAction : public MatchAction { class ConstrainOperandsToDefinitionAction : public MatchAction {
unsigned InsnID; unsigned InsnID;
public: public:
ConstrainOperandsToDefinitionAction(unsigned InsnID) : InsnID(InsnID) {} ConstrainOperandsToDefinitionAction(unsigned InsnID)
: MatchAction(AK_ConstraintOpsToDef), InsnID(InsnID) {}
static bool classof(const MatchAction *A) {
return A->getKind() == AK_ConstraintOpsToDef;
}
void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override { void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override {
Table << MatchTable::Opcode("GIR_ConstrainSelectedInstOperands") Table << MatchTable::Opcode("GIR_ConstrainSelectedInstOperands")
<< MatchTable::Comment("InsnID") << MatchTable::IntValue(InsnID) << MatchTable::Comment("InsnID") << MatchTable::IntValue(InsnID)
<< MatchTable::LineBreak; << MatchTable::LineBreak;
} }
}; };
/// Generates code to constrain the specified operand of an output instruction /// Generates code to constrain the specified operand of an output instruction
/// to the specified register class. /// to the specified register class.
class ConstrainOperandToRegClassAction : public MatchAction { class ConstrainOperandToRegClassAction : public MatchAction {
unsigned InsnID; unsigned InsnID;
unsigned OpIdx; unsigned OpIdx;
const CodeGenRegisterClass &RC; const CodeGenRegisterClass &RC;
public: public:
ConstrainOperandToRegClassAction(unsigned InsnID, unsigned OpIdx, ConstrainOperandToRegClassAction(unsigned InsnID, unsigned OpIdx,
const CodeGenRegisterClass &RC) const CodeGenRegisterClass &RC)
: InsnID(InsnID), OpIdx(OpIdx), RC(RC) {} : MatchAction(AK_ConstraintOpsToRC), InsnID(InsnID), OpIdx(OpIdx),
RC(RC) {}
static bool classof(const MatchAction *A) {
return A->getKind() == AK_ConstraintOpsToRC;
}
void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override; void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override;
}; };
/// Generates code to create a temporary register which can be used to chain /// Generates code to create a temporary register which can be used to chain
/// instructions together. /// instructions together.
class MakeTempRegisterAction : public MatchAction { class MakeTempRegisterAction : public MatchAction {
private: private:
LLTCodeGen Ty; LLTCodeGen Ty;
unsigned TempRegID; unsigned TempRegID;
public: public:
MakeTempRegisterAction(const LLTCodeGen &Ty, unsigned TempRegID) MakeTempRegisterAction(const LLTCodeGen &Ty, unsigned TempRegID)
: Ty(Ty), TempRegID(TempRegID) { : MatchAction(AK_MakeTempReg), Ty(Ty), TempRegID(TempRegID) {
KnownTypes.insert(Ty); KnownTypes.insert(Ty);
} }
static bool classof(const MatchAction *A) {
return A->getKind() == AK_MakeTempReg;
}
void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override; void emitActionOpcodes(MatchTable &Table, RuleMatcher &Rule) const override;
}; };
} // namespace gi } // namespace gi
} // namespace llvm } // namespace llvm
#endif #endif

llvm/utils/TableGen/GlobalISelMatchTable.cpp

Show First 20 Lines • Show All 1,929 Lines • ▼ Show 20 Lines Table << MatchTable::Opcode("GIR_MutateOpcode")
<< MatchTable::NamedValue(I->Namespace, I->TheDef->getName()) << MatchTable::NamedValue(I->Namespace, I->TheDef->getName())
<< MatchTable::LineBreak; << MatchTable::LineBreak;
if (!I->ImplicitDefs.empty() || !I->ImplicitUses.empty()) { if (!I->ImplicitDefs.empty() || !I->ImplicitUses.empty()) {
for (auto *Def : I->ImplicitDefs) { for (auto *Def : I->ImplicitDefs) {
auto Namespace = Def->getValue("Namespace") auto Namespace = Def->getValue("Namespace")
? Def->getValueAsString("Namespace") ? Def->getValueAsString("Namespace")
: ""; : "";
const bool IsDead = DeadImplicitDefs.contains(Def);
Table << MatchTable::Opcode("GIR_AddImplicitDef") Table << MatchTable::Opcode("GIR_AddImplicitDef")
<< MatchTable::Comment("InsnID") << MatchTable::IntValue(InsnID) << MatchTable::Comment("InsnID") << MatchTable::IntValue(InsnID)
<< MatchTable::NamedValue(Namespace, Def->getName()) << MatchTable::NamedValue(Namespace, Def->getName())
<< (IsDead ? MatchTable::NamedValue("RegState", "Dead")
: MatchTable::IntValue(0))
<< MatchTable::LineBreak; << MatchTable::LineBreak;
} }
for (auto *Use : I->ImplicitUses) { for (auto *Use : I->ImplicitUses) {
auto Namespace = Use->getValue("Namespace") auto Namespace = Use->getValue("Namespace")
? Use->getValueAsString("Namespace") ? Use->getValueAsString("Namespace")
: ""; : "";
Table << MatchTable::Opcode("GIR_AddImplicitUse") Table << MatchTable::Opcode("GIR_AddImplicitUse")
<< MatchTable::Comment("InsnID") << MatchTable::IntValue(InsnID) << MatchTable::Comment("InsnID") << MatchTable::IntValue(InsnID)
▲ Show 20 Lines • Show All 71 Lines • Show Last 20 Lines