This is an archive of the discontinued LLVM Phabricator instance.

Differential D24145

AMDGPU: Add runtime metadata for pointee alignment of argument
ClosedPublic

Authored by yaxunl on Sep 1 2016, 10:53 AM.

Details

Reviewers
nhaustov
• tstellarAMD
arsenm
Commits
rGadd05a8d958a: AMDGPU: Add runtime metadata for pointee alignment of argument.
rL280399: AMDGPU: Add runtime metadata for pointee alignment of argument.
Summary

Add runtime metdata for pointee alignment of pointer type kernel argument. The key is KeyArgPointeeAlign and the value is a 32 bit unsigned integer.

Diff Detail

Event Timeline

yaxunl updated this revision to Diff 70027.Sep 1 2016, 10:53 AM
yaxunl retitled this revision from to AMDGPU: Add runtime metadata for pointee alignment of argument.
yaxunl updated this object.
yaxunl added reviewers: arsenm, • tstellarAMD, nhaustov.
yaxunl added subscribers: llvm-commits, Restricted Project.
Herald added subscribers: nhaehnle, wdng, arsenm. · View Herald TranscriptSep 1 2016, 10:53 AM
nhaustov accepted this revision.Sep 1 2016, 11:20 AM
nhaustov edited edge metadata.
This revision is now accepted and ready to land.Sep 1 2016, 11:20 AM
Closed by commit rL280399: AMDGPU: Add runtime metadata for pointee alignment of argument. (authored by yaxunl). · Explain WhySep 1 2016, 11:55 AM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
lib/
Target/
AMDGPU/
8 lines
1 line
test/
CodeGen/
AMDGPU/
167 lines

Diff 70027

lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp

Context not available.
unsigned I = Arg.getArgNo(); unsigned I = Arg.getArgNo();
OutStreamer->EmitIntValue(RuntimeMD::KeyArgBegin, 1); OutStreamer->EmitIntValue(RuntimeMD::KeyArgBegin, 1);
// Emit KeyArgSize and KeyArgAlign. // Emit KeyArgSize, KeyArgAlign and KeyArgPointeeAlign.
Type *T = Arg.getType(); Type *T = Arg.getType();
const DataLayout &DL = F.getParent()->getDataLayout(); const DataLayout &DL = F.getParent()->getDataLayout();
emitRuntimeMDIntValue(*OutStreamer, RuntimeMD::KeyArgSize, emitRuntimeMDIntValue(*OutStreamer, RuntimeMD::KeyArgSize,
DL.getTypeAllocSize(T), 4); DL.getTypeAllocSize(T), 4);
emitRuntimeMDIntValue(*OutStreamer, RuntimeMD::KeyArgAlign, emitRuntimeMDIntValue(*OutStreamer, RuntimeMD::KeyArgAlign,
DL.getABITypeAlignment(T), 4); DL.getABITypeAlignment(T), 4);
if (auto PT = dyn_cast<PointerType>(T)) {
auto ET = PT->getElementType();
if (ET->isSized())
emitRuntimeMDIntValue(*OutStreamer, RuntimeMD::KeyArgPointeeAlign,
DL.getABITypeAlignment(ET), 4);
}
// Emit KeyArgTypeName. // Emit KeyArgTypeName.
auto TypeName = dyn_cast<MDString>(F.getMetadata( auto TypeName = dyn_cast<MDString>(F.getMetadata(
Context not available.

lib/Target/AMDGPU/AMDGPURuntimeMetadata.h

Context not available.
KeyFlatWorkGroupSizeLimits = 29, // Flat work group size limits KeyFlatWorkGroupSizeLimits = 29, // Flat work group size limits
KeyMaxWorkGroupSize = 30, // Maximum work group size KeyMaxWorkGroupSize = 30, // Maximum work group size
KeyNoPartialWorkGroups = 31, // No partial work groups KeyNoPartialWorkGroups = 31, // No partial work groups
KeyArgPointeeAlign = 32, // Alignment of pointee type
}; };
enum Language : uint8_t { enum Language : uint8_t {
Context not available.

test/CodeGen/AMDGPU/runtime-metadata.ll

Context not available.
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 5 ; CHECK-NEXT: .long 5
; CHECK-NEXT: .ascii "int *" ; CHECK-NEXT: .ascii "int *"
Context not available.
; CHECK-NEXT: .long 4 ; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4 ; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .ascii "struct A" ; CHECK-NEXT: .ascii "struct A"
Context not available.
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 5 ; CHECK-NEXT: .long 5
; CHECK-NEXT: .ascii "int *" ; CHECK-NEXT: .ascii "int *"
Context not available.
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 5 ; CHECK-NEXT: .long 5
; CHECK-NEXT: .ascii "int *" ; CHECK-NEXT: .ascii "int *"
Context not available.
; CHECK-NEXT: .long 4 ; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4 ; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 5 ; CHECK-NEXT: .long 5
; CHECK-NEXT: .ascii "int *" ; CHECK-NEXT: .ascii "int *"
Context not available.
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 5 ; CHECK-NEXT: .long 5
; CHECK-NEXT: .ascii "int *" ; CHECK-NEXT: .ascii "int *"
Context not available.
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 5 ; CHECK-NEXT: .long 5
; CHECK-NEXT: .ascii "int *" ; CHECK-NEXT: .ascii "int *"
Context not available.
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 6 ; CHECK-NEXT: .long 6
; CHECK-NEXT: .ascii "int **" ; CHECK-NEXT: .ascii "int **"
Context not available.
; CHECK-NEXT: .long 4 ; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10 ; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4 ; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 11 ; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 8 ; CHECK-NEXT: .long 8
; CHECK-NEXT: .ascii "struct B" ; CHECK-NEXT: .ascii "struct B"
Context not available.
ret void ret void
} }
; CHECK-LABEL:{{^}}test_pointee_align:
; CHECK: .section .AMDGPU.runtime_metadata
; CHECK-NEXT: .byte 4
; CHECK-NEXT: .byte 6
; CHECK-NEXT: .long 18
; CHECK-NEXT: .ascii "test_pointee_align"
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 6
; CHECK-NEXT: .ascii "long *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 9
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 1
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 6
; CHECK-NEXT: .ascii "char *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 1
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 3
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 2
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 7
; CHECK-NEXT: .ascii "char2 *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 1
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 3
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 7
; CHECK-NEXT: .ascii "char3 *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 1
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 3
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 7
; CHECK-NEXT: .ascii "char4 *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 1
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 3
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 8
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 7
; CHECK-NEXT: .ascii "char8 *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 1
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 3
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 7
; CHECK-NEXT: .byte 9
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 10
; CHECK-NEXT: .long 4
; CHECK-NEXT: .byte 32
; CHECK-NEXT: .long 16
; CHECK-NEXT: .byte 11
; CHECK-NEXT: .long 8
; CHECK-NEXT: .ascii "char16 *"
; CHECK-NEXT: .byte 13
; CHECK-NEXT: .byte 1
; CHECK-NEXT: .byte 14
; CHECK-NEXT: .short 1
; CHECK-NEXT: .byte 16
; CHECK-NEXT: .byte 0
; CHECK-NEXT: .byte 15
; CHECK-NEXT: .byte 3
; CHECK-NEXT: .byte 8
; CHECK-NEXT: .byte 5
define amdgpu_kernel void @test_pointee_align(i64 addrspace(1)* %a, i8 addrspace(3)* %b, <2 x i8> addrspace(3)* %c, <3 x i8> addrspace(3)* %d, <4 x i8> addrspace(3)* %e, <8 x i8> addrspace(3)* %f, <16 x i8> addrspace(3)* %g) !kernel_arg_addr_space !91 !kernel_arg_access_qual !92 !kernel_arg_type !93 !kernel_arg_base_type !93 !kernel_arg_type_qual !94 {
ret void
}
!1 = !{i32 0} !1 = !{i32 0}
!2 = !{!"none"} !2 = !{!"none"}
!3 = !{!"int"} !3 = !{!"int"}
Context not available.
!84 = !{!"clk_event_t"} !84 = !{!"clk_event_t"}
!opencl.ocl.version = !{!90} !opencl.ocl.version = !{!90}
!90 = !{i32 2, i32 0} !90 = !{i32 2, i32 0}
!91 = !{i32 0, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3}
!92 = !{!"none", !"none", !"none", !"none", !"none", !"none", !"none"}
!93 = !{!"long *", !"char *", !"char2 *", !"char3 *", !"char4 *", !"char8 *", !"char16 *"}
!94 = !{!"", !"", !"", !"", !"", !"", !""}
Context not available.