This is an archive of the discontinued LLVM Phabricator instance.

Differential D139829

[AMDGPU] Make use of !listremove. NFCI.
ClosedPublic

Authored by foad on Dec 12 2022, 3:32 AM.

Details

Reviewers
Joe_Nash
RKSimon
arsenm
Group Reviewers
Restricted Project
Commits
rG81084bfa2c3c: [AMDGPU] Make use of !listremove. NFCI.
Summary

This only affects the order of implicit operands in some MIR tests.

Diff Detail

Event Timeline

foad created this revision.Dec 12 2022, 3:32 AM
Herald added a project: Restricted Project. · View Herald TranscriptDec 12 2022, 3:32 AM
Herald added subscribers: kosarev, kerbowa, hiraditya and 6 others. · View Herald Transcript
foad requested review of this revision.Dec 12 2022, 3:32 AM
Herald added a project: Restricted Project. · View Herald TranscriptDec 12 2022, 3:32 AM
Herald added subscribers: llvm-commits, wdng. · View Herald Transcript
Harbormaster completed remote builds in B202536: Diff 482058.Dec 12 2022, 4:39 AM
RKSimon added a comment.Dec 12 2022, 7:53 AM

Nice use case!

arsenm accepted this revision.Dec 12 2022, 7:54 AM
arsenm added inline comments.
llvm/test/CodeGen/AMDGPU/vopd-combine.mir
41

I did want to try to consistently keep $exec last

This revision is now accepted and ready to land.Dec 12 2022, 7:54 AM
Joe_Nash accepted this revision.Dec 12 2022, 7:57 AM

LGTM

Closed by commit rG81084bfa2c3c: [AMDGPU] Make use of !listremove. NFCI. (authored by foad). · Explain WhyDec 12 2022, 9:02 AM
This revision was automatically updated to reflect the committed changes.
foad added a commit: rG81084bfa2c3c: [AMDGPU] Make use of !listremove. NFCI..

Revision Contents

PathSize
llvm/
lib/
Target/
AMDGPU/
11 lines
test/
CodeGen/
AMDGPU/
34 lines
8 lines

Diff 482153

llvm/lib/Target/AMDGPU/AMDGPUInstructions.td

Show First 20 Lines • Show All 77 Lines • ▼ Show 20 Lines
def TruePredicate : Predicate<"">; def TruePredicate : Predicate<"">;
// FIXME: Tablegen should specially supports this // FIXME: Tablegen should specially supports this
def FalsePredicate : Predicate<"false">; def FalsePredicate : Predicate<"false">;
// Add a predicate to the list if does not already exist to deduplicate it. // Add a predicate to the list if does not already exist to deduplicate it.
class PredConcat<list<Predicate> lst, Predicate pred> { class PredConcat<list<Predicate> lst, Predicate pred> {
list<Predicate> ret = list<Predicate> ret = !listconcat(lst, !listremove([pred], lst));
!listconcat([pred], !filter(item, lst, !ne(item, pred)));
} }
// Add a Register to the list if does not already exist
class RegAppend<list<Register> lst, Register reg> {
list<Register> ret =
!listconcat([reg], !filter(item, lst, !ne(item, reg)));
}
// Get the union of two Register lists // Get the union of two Register lists
class RegListUnion<list<Register> lstA, list<Register> lstB> { class RegListUnion<list<Register> lstA, list<Register> lstB> {
list<Register> ret = list<Register> ret = !listconcat(lstA, !listremove(lstB, lstA));
!foldl(lstA, lstB, temp, item, RegAppend<temp, item>.ret);
} }
class PredicateControl { class PredicateControl {
Predicate SubtargetPredicate = TruePredicate; Predicate SubtargetPredicate = TruePredicate;
Predicate AssemblerPredicate = TruePredicate; Predicate AssemblerPredicate = TruePredicate;
Predicate WaveSizePredicate = TruePredicate; Predicate WaveSizePredicate = TruePredicate;
list<Predicate> OtherPredicates = []; list<Predicate> OtherPredicates = [];
list<Predicate> Predicates = PredConcat< list<Predicate> Predicates = PredConcat<
▲ Show 20 Lines • Show All 743 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/vopd-combine.mir

Show All 32 Lines bb.0:
; SCHED-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr3 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr3 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr6 = V_MUL_F32_e32 killed $vgpr0, $vgpr0, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr6 = V_MUL_F32_e32 killed $vgpr0, $vgpr0, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr4 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr4 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_schedule ; PAIR-LABEL: name: vopd_schedule
; PAIR: $vgpr0 = IMPLICIT_DEF ; PAIR: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, killed $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, killed $vgpr0, $vgpr0, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
arsenmUnsubmitted
Not Done
ReplyInline Actions

I did want to try to consistently keep $exec last

arsenm: I did want to try to consistently keep $exec last
; PAIR-NEXT: $vgpr4 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr4 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; can fuse vgpr3 and vgpr6 writing insts only due to reg constraints ; can fuse vgpr3 and vgpr6 writing insts only due to reg constraints
$vgpr3 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec $vgpr3 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
$vgpr4 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec $vgpr4 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
$vgpr6 = V_MUL_F32_e32 $vgpr0, $vgpr0, implicit $mode, implicit $exec $vgpr6 = V_MUL_F32_e32 $vgpr0, $vgpr0, implicit $mode, implicit $exec
Show All 13 Lines bb.0:
; SCHED-NEXT: $vgpr3 = IMPLICIT_DEF ; SCHED-NEXT: $vgpr3 = IMPLICIT_DEF
; SCHED-NEXT: $vgpr5 = V_FMAMK_F32 killed $vgpr0, 10, killed $vgpr3, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr5 = V_FMAMK_F32 killed $vgpr0, 10, killed $vgpr3, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr2 = V_FMAC_F32_e32 killed $vgpr1, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr2 = V_FMAC_F32_e32 killed $vgpr1, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_fmamk ; PAIR-LABEL: name: vopd_fmamk
; PAIR: $vgpr2 = IMPLICIT_DEF ; PAIR: $vgpr2 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr5, $vgpr2 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32 killed $vgpr0, 10, killed $vgpr3, killed $vgpr1, $vgpr1, killed $vgpr2, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr5, $vgpr2 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32 killed $vgpr0, 10, killed $vgpr3, killed $vgpr1, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
; should pair ; should pair
$vgpr2 = V_FMAC_F32_e32 $vgpr1, $vgpr1, $vgpr2, implicit $mode, implicit $exec $vgpr2 = V_FMAC_F32_e32 $vgpr1, $vgpr1, $vgpr2, implicit $mode, implicit $exec
$vgpr5 = V_FMAMK_F32 $vgpr0, 10, $vgpr3, implicit $mode, implicit $exec $vgpr5 = V_FMAMK_F32 $vgpr0, 10, $vgpr3, implicit $mode, implicit $exec
▲ Show 20 Lines • Show All 48 Lines • ▼ Show 20 Lines bb.0:
; SCHED-NEXT: $vgpr9 = V_CNDMASK_B32_e32 killed $sgpr20, killed $vgpr3, implicit $mode, implicit $exec, implicit $vcc ; SCHED-NEXT: $vgpr9 = V_CNDMASK_B32_e32 killed $sgpr20, killed $vgpr3, implicit $mode, implicit $exec, implicit $vcc
; PAIR-LABEL: name: vopd_cndmask ; PAIR-LABEL: name: vopd_cndmask
; PAIR: $vgpr2 = IMPLICIT_DEF ; PAIR: $vgpr2 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF
; PAIR-NEXT: $sgpr20 = IMPLICIT_DEF ; PAIR-NEXT: $sgpr20 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr4 = V_FMAMK_F32 $sgpr20, 12345, $vgpr3, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr4 = V_FMAMK_F32 $sgpr20, 12345, $vgpr3, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr2, $vgpr5 = V_DUAL_FMAC_F32_e32_X_CNDMASK_B32_e32 $sgpr20, killed $vgpr1, killed $vgpr2, $vgpr0, $vgpr3, implicit $exec, implicit $vcc, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr2, $vgpr5 = V_DUAL_FMAC_F32_e32_X_CNDMASK_B32_e32 $sgpr20, killed $vgpr1, killed $vgpr2, $vgpr0, $vgpr3, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr7 = V_CNDMASK_B32_e32 killed $vgpr0, $vgpr3, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr7 = V_CNDMASK_B32_e32 killed $vgpr0, $vgpr3, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr6 = V_ADD_F32_e32 $sgpr20, $vgpr3, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr6 = V_ADD_F32_e32 $sgpr20, $vgpr3, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr9 = V_CNDMASK_B32_e32 killed $sgpr20, killed $vgpr3, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr9 = V_CNDMASK_B32_e32 killed $sgpr20, killed $vgpr3, implicit $mode, implicit $exec, implicit $vcc
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
$sgpr20 = IMPLICIT_DEF $sgpr20 = IMPLICIT_DEF
▲ Show 20 Lines • Show All 96 Lines • ▼ Show 20 Lines bb.0:
; SCHED-NEXT: $vgpr3 = IMPLICIT_DEF ; SCHED-NEXT: $vgpr3 = IMPLICIT_DEF
; SCHED-NEXT: $vgpr5 = V_FMAMK_F32 killed $vgpr0, 100, killed $vgpr3, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr5 = V_FMAMK_F32 killed $vgpr0, 100, killed $vgpr3, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr2 = V_FMAC_F32_e32 4, killed $vgpr1, killed $vgpr2, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr2 = V_FMAC_F32_e32 4, killed $vgpr1, killed $vgpr2, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_constants_inlinable ; PAIR-LABEL: name: vopd_constants_inlinable
; PAIR: $vgpr2 = IMPLICIT_DEF ; PAIR: $vgpr2 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr5, $vgpr2 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32 killed $vgpr0, 100, killed $vgpr3, 4, killed $vgpr1, killed $vgpr2, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr5, $vgpr2 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32 killed $vgpr0, 100, killed $vgpr3, 4, killed $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
; can pair since 4 is inlinable ; can pair since 4 is inlinable
$vgpr2 = V_FMAC_F32_e32 4, $vgpr1, $vgpr2, implicit $mode, implicit $exec $vgpr2 = V_FMAC_F32_e32 4, $vgpr1, $vgpr2, implicit $mode, implicit $exec
$vgpr5 = V_FMAMK_F32 $vgpr0, 100, $vgpr3, implicit $mode, implicit $exec $vgpr5 = V_FMAMK_F32 $vgpr0, 100, $vgpr3, implicit $mode, implicit $exec
Show All 13 Lines bb.0:
; SCHED-NEXT: $vgpr3 = IMPLICIT_DEF ; SCHED-NEXT: $vgpr3 = IMPLICIT_DEF
; SCHED-NEXT: $vgpr5 = V_FMAMK_F32 killed $vgpr0, 100, killed $vgpr3, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr5 = V_FMAMK_F32 killed $vgpr0, 100, killed $vgpr3, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr2 = V_FMAC_F32_e32 100, killed $vgpr1, killed $vgpr2, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr2 = V_FMAC_F32_e32 100, killed $vgpr1, killed $vgpr2, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_constants_same ; PAIR-LABEL: name: vopd_constants_same
; PAIR: $vgpr2 = IMPLICIT_DEF ; PAIR: $vgpr2 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr5, $vgpr2 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32 killed $vgpr0, 100, killed $vgpr3, 100, killed $vgpr1, killed $vgpr2, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr5, $vgpr2 = V_DUAL_FMAMK_F32_X_FMAC_F32_e32 killed $vgpr0, 100, killed $vgpr3, 100, killed $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
; should be able to pair using 1 deduplicated literal ; should be able to pair using 1 deduplicated literal
$vgpr2 = V_FMAC_F32_e32 100, $vgpr1, $vgpr2, implicit $mode, implicit $exec $vgpr2 = V_FMAC_F32_e32 100, $vgpr1, $vgpr2, implicit $mode, implicit $exec
$vgpr5 = V_FMAMK_F32 $vgpr0, 100, $vgpr3, implicit $mode, implicit $exec $vgpr5 = V_FMAMK_F32 $vgpr0, 100, $vgpr3, implicit $mode, implicit $exec
Show All 32 Lines bb.0:
; SCHED: $vgpr0 = IMPLICIT_DEF ; SCHED: $vgpr0 = IMPLICIT_DEF
; SCHED-NEXT: $vgpr1 = IMPLICIT_DEF ; SCHED-NEXT: $vgpr1 = IMPLICIT_DEF
; SCHED-NEXT: $vgpr3 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr3 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
; SCHED-NEXT: DBG_VALUE $vgpr0, 0, 0 ; SCHED-NEXT: DBG_VALUE $vgpr0, 0, 0
; SCHED-NEXT: $vgpr6 = V_MUL_F32_e32 killed $vgpr0, $vgpr0, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr6 = V_MUL_F32_e32 killed $vgpr0, $vgpr0, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_debug ; PAIR-LABEL: name: vopd_debug
; PAIR: $vgpr0 = IMPLICIT_DEF ; PAIR: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 killed $vgpr1, $vgpr1, killed $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 killed $vgpr1, $vgpr1, killed $vgpr0, $vgpr0, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; PAIR-NEXT: DBG_VALUE $vgpr0, 0, 0 ; PAIR-NEXT: DBG_VALUE $vgpr0, 0, 0
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
; TODO Debug values disable VOPD creation ; TODO Debug values disable VOPD creation
$vgpr3 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec $vgpr3 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
DBG_VALUE $vgpr0, 0, 0 DBG_VALUE $vgpr0, 0, 0
$vgpr6 = V_MUL_F32_e32 $vgpr0, $vgpr0, implicit $mode, implicit $exec $vgpr6 = V_MUL_F32_e32 $vgpr0, $vgpr0, implicit $mode, implicit $exec
Show All 24 Lines bb.0:
; SCHED-NEXT: $vgpr16 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr16 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr14 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr14 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_schedule_unconstrained ; PAIR-LABEL: name: vopd_schedule_unconstrained
; PAIR: $vgpr2 = IMPLICIT_DEF ; PAIR: $vgpr2 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr4 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr4 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr0, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr2 = V_FMAC_F32_e32 10, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr2 = V_FMAC_F32_e32 10, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr12, $vgpr19 = V_DUAL_ADD_F32_e32_X_CNDMASK_B32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr3, implicit $exec, implicit $vcc, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr12, $vgpr19 = V_DUAL_ADD_F32_e32_X_CNDMASK_B32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr3, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr11 = V_CNDMASK_B32_e32 $vgpr0, killed $vgpr3, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr11 = V_CNDMASK_B32_e32 $vgpr0, killed $vgpr3, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr17, $vgpr10 = V_DUAL_MUL_F32_e32_X_CNDMASK_B32_e32 killed $vgpr0, $vgpr0, $vgpr1, $vgpr2, implicit $exec, implicit $vcc, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr17, $vgpr10 = V_DUAL_MUL_F32_e32_X_CNDMASK_B32_e32 killed $vgpr0, $vgpr0, $vgpr1, $vgpr2, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr15 = V_CNDMASK_B32_e32 $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr15 = V_CNDMASK_B32_e32 $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr16 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr16 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr14 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr14 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
$vgpr2 = V_FMAC_F32_e32 10, $vgpr1, $vgpr2, implicit $mode, implicit $exec $vgpr2 = V_FMAC_F32_e32 10, $vgpr1, $vgpr2, implicit $mode, implicit $exec
▲ Show 20 Lines • Show All 54 Lines • ▼ Show 20 Lines bb.0:
; SCHED-NEXT: $vgpr34 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr34 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; SCHED-NEXT: $vgpr32 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; SCHED-NEXT: $vgpr32 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-LABEL: name: vopd_schedule_unconstrained_2 ; PAIR-LABEL: name: vopd_schedule_unconstrained_2
; PAIR: $vgpr2 = IMPLICIT_DEF ; PAIR: $vgpr2 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr3 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr0 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr1 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr20 = IMPLICIT_DEF ; PAIR-NEXT: $vgpr20 = IMPLICIT_DEF
; PAIR-NEXT: $vgpr16, $vgpr35 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr16, $vgpr35 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr0, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr3, $vgpr6 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr0, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr2 = V_FMAC_F32_e32 10, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr2 = V_FMAC_F32_e32 10, $vgpr1, killed $vgpr2, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr2 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr4, $vgpr29 = V_DUAL_SUB_F32_e32_X_CNDMASK_B32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr3, implicit $exec, implicit $vcc, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr4, $vgpr29 = V_DUAL_SUB_F32_e32_X_CNDMASK_B32_e32 $vgpr1, $vgpr1, $vgpr0, $vgpr3, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr19, $vgpr20 = V_DUAL_CNDMASK_B32_e32_X_FMAC_F32_e32 $vgpr0, $vgpr3, 10, $vgpr1, killed $vgpr20, implicit $exec, implicit $mode, implicit $vcc, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr19, $vgpr20 = V_DUAL_CNDMASK_B32_e32_X_FMAC_F32_e32 $vgpr0, $vgpr3, 10, $vgpr1, killed $vgpr20, implicit $vcc, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr15 = V_CNDMASK_B32_e32 $vgpr1, $vgpr2, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr15 = V_CNDMASK_B32_e32 $vgpr1, $vgpr2, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr10, $vgpr17 = V_DUAL_CNDMASK_B32_e32_X_MUL_F32_e32 $vgpr1, $vgpr2, $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $vcc, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr10, $vgpr17 = V_DUAL_CNDMASK_B32_e32_X_MUL_F32_e32 $vgpr1, $vgpr2, $vgpr0, $vgpr0, implicit $vcc, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr11, $vgpr12 = V_DUAL_CNDMASK_B32_e32_X_ADD_F32_e32 $vgpr0, $vgpr3, $vgpr1, $vgpr1, implicit $exec, implicit $mode, implicit $vcc, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr11, $vgpr12 = V_DUAL_CNDMASK_B32_e32_X_ADD_F32_e32 $vgpr0, $vgpr3, $vgpr1, $vgpr1, implicit $vcc, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr37, $vgpr14 = V_DUAL_CNDMASK_B32_e32_X_SUB_F32_e32 $vgpr0, killed $vgpr3, $vgpr1, $vgpr1, implicit $exec, implicit $mode, implicit $vcc, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr37, $vgpr14 = V_DUAL_CNDMASK_B32_e32_X_SUB_F32_e32 $vgpr0, killed $vgpr3, $vgpr1, $vgpr1, implicit $vcc, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $vcc, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr20 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr20 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr21, $vgpr24 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, killed $vgpr0, $vgpr0, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr21, $vgpr24 = V_DUAL_SUB_F32_e32_X_MUL_F32_e32 $vgpr1, $vgpr1, killed $vgpr0, $vgpr0, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr28 = V_CNDMASK_B32_e32 $vgpr1, $vgpr2, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr28 = V_CNDMASK_B32_e32 $vgpr1, $vgpr2, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr22 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr22 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr31 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr31 = V_ADD_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr33 = V_CNDMASK_B32_e32 $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $vcc ; PAIR-NEXT: $vgpr33 = V_CNDMASK_B32_e32 $vgpr1, killed $vgpr2, implicit $mode, implicit $exec, implicit $vcc
; PAIR-NEXT: $vgpr34 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr34 = V_SUB_F32_e32 $vgpr1, $vgpr1, implicit $mode, implicit $exec
; PAIR-NEXT: $vgpr32 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec ; PAIR-NEXT: $vgpr32 = V_SUB_F32_e32 killed $vgpr1, $vgpr1, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
▲ Show 20 Lines • Show All 79 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/vopd-src2acc-delay.mir

# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -march=amdgcn -mcpu=gfx1100 -verify-machineinstrs -run-pass=gcn-create-vopd,amdgpu-insert-delay-alu %s -o - | FileCheck %s # RUN: llc -march=amdgcn -mcpu=gfx1100 -verify-machineinstrs -run-pass=gcn-create-vopd,amdgpu-insert-delay-alu %s -o - | FileCheck %s
--- ---
name: vopd_fmac_fmac name: vopd_fmac_fmac
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
; CHECK-LABEL: name: vopd_fmac_fmac ; CHECK-LABEL: name: vopd_fmac_fmac
; CHECK: $vgpr0 = IMPLICIT_DEF ; CHECK: $vgpr0 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr1 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr1 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr2 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr2 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr3 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr3 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr4 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr4 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_FMAC_F32_e32_X_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_FMAC_F32_e32_X_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; CHECK-NEXT: S_DELAY_ALU 1 ; CHECK-NEXT: S_DELAY_ALU 1
; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_FMAC_F32_e32_X_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_FMAC_F32_e32_X_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
$vgpr4 = IMPLICIT_DEF $vgpr4 = IMPLICIT_DEF
$vgpr0 = V_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec $vgpr0 = V_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec
$vgpr1 = V_FMAC_F32_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec $vgpr1 = V_FMAC_F32_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec
$vgpr0 = V_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec $vgpr0 = V_FMAC_F32_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec
$vgpr1 = V_FMAC_F32_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec $vgpr1 = V_FMAC_F32_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec
... ...
--- ---
name: vopd_dot2c_dot2c name: vopd_dot2c_dot2c
tracksRegLiveness: true tracksRegLiveness: true
body: | body: |
bb.0: bb.0:
; CHECK-LABEL: name: vopd_dot2c_dot2c ; CHECK-LABEL: name: vopd_dot2c_dot2c
; CHECK: $vgpr0 = IMPLICIT_DEF ; CHECK: $vgpr0 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr1 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr1 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr2 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr2 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr3 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr3 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr4 = IMPLICIT_DEF ; CHECK-NEXT: $vgpr4 = IMPLICIT_DEF
; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_DOT2C_F32_F16_e32_X_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_DOT2C_F32_F16_e32_X_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
; CHECK-NEXT: S_DELAY_ALU 1 ; CHECK-NEXT: S_DELAY_ALU 1
; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_DOT2C_F32_F16_e32_X_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $exec, implicit $mode, implicit $mode, implicit $exec, implicit $mode, implicit $exec ; CHECK-NEXT: $vgpr0, $vgpr1 = V_DUAL_DOT2C_F32_F16_e32_X_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec, implicit $mode, implicit $exec, implicit $mode, implicit $exec
$vgpr0 = IMPLICIT_DEF $vgpr0 = IMPLICIT_DEF
$vgpr1 = IMPLICIT_DEF $vgpr1 = IMPLICIT_DEF
$vgpr2 = IMPLICIT_DEF $vgpr2 = IMPLICIT_DEF
$vgpr3 = IMPLICIT_DEF $vgpr3 = IMPLICIT_DEF
$vgpr4 = IMPLICIT_DEF $vgpr4 = IMPLICIT_DEF
$vgpr0 = V_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec $vgpr0 = V_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec
$vgpr1 = V_DOT2C_F32_F16_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec $vgpr1 = V_DOT2C_F32_F16_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec
$vgpr0 = V_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec $vgpr0 = V_DOT2C_F32_F16_e32 $vgpr2, $vgpr3, $vgpr0, implicit $mode, implicit $exec
$vgpr1 = V_DOT2C_F32_F16_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec $vgpr1 = V_DOT2C_F32_F16_e32 $vgpr3, $vgpr4, $vgpr1, implicit $mode, implicit $exec
... ...