Skip to content

[AMDGPU] Extend ComputePGMRSrc3 to gfx10+. NFCI. #129289

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Mar 3, 2025
Merged

[AMDGPU] Extend ComputePGMRSrc3 to gfx10+. NFCI. #129289

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 13 additions & 12 deletions llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -622,12 +622,13 @@ AMDGPUAsmPrinter::getAmdhsaKernelDescriptor(const MachineFunction &MF,

int64_t PGRM_Rsrc3 = 1;
bool EvaluatableRsrc3 =
CurrentProgramInfo.ComputePGMRSrc3GFX90A->evaluateAsAbsolute(PGRM_Rsrc3);
CurrentProgramInfo.ComputePGMRSrc3->evaluateAsAbsolute(PGRM_Rsrc3);
(void)PGRM_Rsrc3;
(void)EvaluatableRsrc3;
assert(STM.hasGFX90AInsts() || !EvaluatableRsrc3 ||
assert(STM.getGeneration() >= AMDGPUSubtarget::GFX10 ||
STM.hasGFX90AInsts() || !EvaluatableRsrc3 ||
static_cast<uint64_t>(PGRM_Rsrc3) == 0);
KernelDescriptor.compute_pgm_rsrc3 = CurrentProgramInfo.ComputePGMRSrc3GFX90A;
KernelDescriptor.compute_pgm_rsrc3 = CurrentProgramInfo.ComputePGMRSrc3;

KernelDescriptor.kernarg_preload = MCConstantExpr::create(
AMDGPU::hasKernargPreload(STM) ? Info->getNumKernargPreloadedSGPRs() : 0,
Expand Down Expand Up @@ -822,22 +823,22 @@ bool AMDGPUAsmPrinter::runOnMachineFunction(MachineFunction &MF) {
false);

[[maybe_unused]] int64_t PGMRSrc3;
assert(STM.hasGFX90AInsts() ||
(CurrentProgramInfo.ComputePGMRSrc3GFX90A->evaluateAsAbsolute(
PGMRSrc3) &&
assert(STM.getGeneration() >= AMDGPUSubtarget::GFX10 ||
STM.hasGFX90AInsts() ||
(CurrentProgramInfo.ComputePGMRSrc3->evaluateAsAbsolute(PGMRSrc3) &&
static_cast<uint64_t>(PGMRSrc3) == 0));
if (STM.hasGFX90AInsts()) {
OutStreamer->emitRawComment(
" COMPUTE_PGM_RSRC3_GFX90A:ACCUM_OFFSET: " +
getMCExprStr(MCKernelDescriptor::bits_get(
CurrentProgramInfo.ComputePGMRSrc3GFX90A,
CurrentProgramInfo.ComputePGMRSrc3,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_ACCUM_OFFSET_SHIFT,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_ACCUM_OFFSET, Ctx)),
false);
OutStreamer->emitRawComment(
" COMPUTE_PGM_RSRC3_GFX90A:TG_SPLIT: " +
getMCExprStr(MCKernelDescriptor::bits_get(
CurrentProgramInfo.ComputePGMRSrc3GFX90A,
CurrentProgramInfo.ComputePGMRSrc3,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT_SHIFT,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT, Ctx)),
false);
Expand Down Expand Up @@ -1241,12 +1242,12 @@ void AMDGPUAsmPrinter::getSIProgramInfo(SIProgramInfo &ProgInfo,
return Dst;
};

ProgInfo.ComputePGMRSrc3GFX90A =
SetBits(ProgInfo.ComputePGMRSrc3GFX90A, ProgInfo.AccumOffset,
ProgInfo.ComputePGMRSrc3 =
SetBits(ProgInfo.ComputePGMRSrc3, ProgInfo.AccumOffset,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_ACCUM_OFFSET,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_ACCUM_OFFSET_SHIFT);
ProgInfo.ComputePGMRSrc3GFX90A =
SetBits(ProgInfo.ComputePGMRSrc3GFX90A, CreateExpr(ProgInfo.TgSplit),
ProgInfo.ComputePGMRSrc3 =
SetBits(ProgInfo.ComputePGMRSrc3, CreateExpr(ProgInfo.TgSplit),
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT,
amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT_SHIFT);
}
Expand Down
2 changes: 1 addition & 1 deletion llvm/lib/Target/AMDGPU/SIProgramInfo.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,7 +57,7 @@ void SIProgramInfo::reset(const MachineFunction &MF) {
LdsSize = 0;
EXCPEnable = 0;

ComputePGMRSrc3GFX90A = ZeroExpr;
ComputePGMRSrc3 = ZeroExpr;

NumVGPR = ZeroExpr;
NumArchVGPR = ZeroExpr;
Expand Down
2 changes: 1 addition & 1 deletion llvm/lib/Target/AMDGPU/SIProgramInfo.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,7 @@ struct LLVM_EXTERNAL_VISIBILITY SIProgramInfo {
uint32_t LdsSize = 0;
uint32_t EXCPEnable = 0;

const MCExpr *ComputePGMRSrc3GFX90A = nullptr;
const MCExpr *ComputePGMRSrc3 = nullptr;

const MCExpr *NumVGPR = nullptr;
const MCExpr *NumArchVGPR = nullptr;
Expand Down