[AMDGPU] Support SIProgramInfo MCExpr for comments and remarks #94350
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-backend-amdgpu Author: Janek van Oirschot (JanekvO) ChangesEliminates assumption that MCExpr comments/remarks being emitted are always resolvable Full diff: https://github.com/llvm/llvm-project/pull/94350.diff 2 Files Affected:
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
index cad4a3430327b..7ac2f9db5e046 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
@@ -400,6 +400,37 @@ void AMDGPUAsmPrinter::emitCommonFunctionComments(
false);
}
+std::string AMDGPUAsmPrinter::getMCExprStr(const MCExpr *Value) {
+ std::string Str;
+ raw_string_ostream OSS(Str);
+ int64_t IVal;
+ if (Value->evaluateAsAbsolute(IVal)) {
+ OSS << static_cast<uint64_t>(IVal);
+ } else {
+ Value->print(OSS, MAI);
+ }
+ return Str;
+}
+
+void AMDGPUAsmPrinter::emitCommonFunctionComments(
+ const MCExpr *NumVGPR, std::optional<const MCExpr *> NumAGPR,
+ const MCExpr *TotalNumVGPR, const MCExpr *NumSGPR,
+ const MCExpr *ScratchSize, uint64_t CodeSize,
+ const AMDGPUMachineFunction *MFI) {
+ OutStreamer->emitRawComment(" codeLenInByte = " + Twine(CodeSize), false);
+ OutStreamer->emitRawComment(" NumSgprs: " + getMCExprStr(NumSGPR), false);
+ OutStreamer->emitRawComment(" NumVgprs: " + getMCExprStr(NumVGPR), false);
+ if (NumAGPR) {
+ OutStreamer->emitRawComment(" NumAgprs: " + getMCExprStr(*NumAGPR), false);
+ OutStreamer->emitRawComment(" TotalNumVgprs: " + getMCExprStr(TotalNumVGPR),
+ false);
+ }
+ OutStreamer->emitRawComment(" ScratchSize: " + getMCExprStr(ScratchSize),
+ false);
+ OutStreamer->emitRawComment(" MemoryBound: " + Twine(MFI->isMemoryBound()),
+ false);
+}
+
uint16_t AMDGPUAsmPrinter::getAmdhsaKernelCodeProperties(
const MachineFunction &MF) const {
const SIMachineFunctionInfo &MFI = *MF.getInfo<SIMachineFunctionInfo>();
@@ -554,13 +585,11 @@ bool AMDGPUAsmPrinter::runOnMachineFunction(MachineFunction &MF) {
OutStreamer->emitRawComment(" Kernel info:", false);
emitCommonFunctionComments(
- getMCExprValue(CurrentProgramInfo.NumArchVGPR, Ctx),
- STM.hasMAIInsts() ? getMCExprValue(CurrentProgramInfo.NumAccVGPR, Ctx)
- : std::optional<uint32_t>(),
- getMCExprValue(CurrentProgramInfo.NumVGPR, Ctx),
- getMCExprValue(CurrentProgramInfo.NumSGPR, Ctx),
- getMCExprValue(CurrentProgramInfo.ScratchSize, Ctx),
- getFunctionCodeSize(MF), MFI);
+ CurrentProgramInfo.NumArchVGPR,
+ STM.hasMAIInsts() ? CurrentProgramInfo.NumAccVGPR
+ : std::optional<const MCExpr *>(),
+ CurrentProgramInfo.NumVGPR, CurrentProgramInfo.NumSGPR,
+ CurrentProgramInfo.ScratchSize, getFunctionCodeSize(MF), MFI);
OutStreamer->emitRawComment(
" FloatMode: " + Twine(CurrentProgramInfo.FloatMode), false);
@@ -571,43 +600,38 @@ bool AMDGPUAsmPrinter::runOnMachineFunction(MachineFunction &MF) {
" bytes/workgroup (compile time only)", false);
OutStreamer->emitRawComment(
- " SGPRBlocks: " +
- Twine(getMCExprValue(CurrentProgramInfo.SGPRBlocks, Ctx)),
- false);
+ " SGPRBlocks: " + getMCExprStr(CurrentProgramInfo.SGPRBlocks), false);
+
OutStreamer->emitRawComment(
- " VGPRBlocks: " +
- Twine(getMCExprValue(CurrentProgramInfo.VGPRBlocks, Ctx)),
- false);
+ " VGPRBlocks: " + getMCExprStr(CurrentProgramInfo.VGPRBlocks), false);
OutStreamer->emitRawComment(
" NumSGPRsForWavesPerEU: " +
- Twine(
- getMCExprValue(CurrentProgramInfo.NumSGPRsForWavesPerEU, Ctx)),
+ getMCExprStr(CurrentProgramInfo.NumSGPRsForWavesPerEU),
false);
OutStreamer->emitRawComment(
" NumVGPRsForWavesPerEU: " +
- Twine(
- getMCExprValue(CurrentProgramInfo.NumVGPRsForWavesPerEU, Ctx)),
+ getMCExprStr(CurrentProgramInfo.NumVGPRsForWavesPerEU),
false);
- if (STM.hasGFX90AInsts())
+ if (STM.hasGFX90AInsts()) {
+ const MCExpr *AdjustedAccum = MCBinaryExpr::createAdd(
+ CurrentProgramInfo.AccumOffset, MCConstantExpr::create(1, Ctx), Ctx);
+ AdjustedAccum = MCBinaryExpr::createMul(
+ AdjustedAccum, MCConstantExpr::create(4, Ctx), Ctx);
OutStreamer->emitRawComment(
- " AccumOffset: " +
- Twine((getMCExprValue(CurrentProgramInfo.AccumOffset, Ctx) + 1) *
- 4),
- false);
+ " AccumOffset: " + getMCExprStr(AdjustedAccum), false);
+ }
OutStreamer->emitRawComment(
- " Occupancy: " +
- Twine(getMCExprValue(CurrentProgramInfo.Occupancy, Ctx)),
- false);
+ " Occupancy: " + getMCExprStr(CurrentProgramInfo.Occupancy), false);
OutStreamer->emitRawComment(
" WaveLimiterHint : " + Twine(MFI->needsWaveLimiter()), false);
OutStreamer->emitRawComment(
" COMPUTE_PGM_RSRC2:SCRATCH_EN: " +
- Twine(getMCExprValue(CurrentProgramInfo.ScratchEnable, Ctx)),
+ getMCExprStr(CurrentProgramInfo.ScratchEnable),
false);
OutStreamer->emitRawComment(" COMPUTE_PGM_RSRC2:USER_SGPR: " +
Twine(CurrentProgramInfo.UserSGPR),
@@ -628,20 +652,25 @@ bool AMDGPUAsmPrinter::runOnMachineFunction(MachineFunction &MF) {
Twine(CurrentProgramInfo.TIdIGCompCount),
false);
+ int64_t PGMRSrc3;
assert(STM.hasGFX90AInsts() ||
- getMCExprValue(CurrentProgramInfo.ComputePGMRSrc3GFX90A, Ctx) == 0);
+ (CurrentProgramInfo.ComputePGMRSrc3GFX90A->evaluateAsAbsolute(
+ PGMRSrc3) &&
+ static_cast<uint64_t>(PGMRSrc3) == 0));
if (STM.hasGFX90AInsts()) {
OutStreamer->emitRawComment(
" COMPUTE_PGM_RSRC3_GFX90A:ACCUM_OFFSET: " +
- Twine((AMDHSA_BITS_GET(
- getMCExprValue(CurrentProgramInfo.ComputePGMRSrc3GFX90A, Ctx),
- amdhsa::COMPUTE_PGM_RSRC3_GFX90A_ACCUM_OFFSET))),
+ getMCExprStr(MCKernelDescriptor::bits_get(
+ CurrentProgramInfo.ComputePGMRSrc3GFX90A,
+ 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: " +
- Twine((AMDHSA_BITS_GET(
- getMCExprValue(CurrentProgramInfo.ComputePGMRSrc3GFX90A, Ctx),
- amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT))),
+ getMCExprStr(MCKernelDescriptor::bits_get(
+ CurrentProgramInfo.ComputePGMRSrc3GFX90A,
+ amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT_SHIFT,
+ amdhsa::COMPUTE_PGM_RSRC3_GFX90A_TG_SPLIT, Ctx)),
false);
}
}
@@ -1463,28 +1492,26 @@ void AMDGPUAsmPrinter::emitResourceUsageRemarks(
// remarks to simulate newlines. If and when clang does accept newlines, this
// formatting should be aggregated into one remark with newlines to avoid
// printing multiple diagnostic location and diag opts.
- MCContext &MCCtx = MF.getContext();
EmitResourceUsageRemark("FunctionName", "Function Name",
MF.getFunction().getName());
EmitResourceUsageRemark("NumSGPR", "SGPRs",
- getMCExprValue(CurrentProgramInfo.NumSGPR, MCCtx));
- EmitResourceUsageRemark(
- "NumVGPR", "VGPRs",
- getMCExprValue(CurrentProgramInfo.NumArchVGPR, MCCtx));
+ getMCExprStr(CurrentProgramInfo.NumSGPR));
+ EmitResourceUsageRemark("NumVGPR", "VGPRs",
+ getMCExprStr(CurrentProgramInfo.NumArchVGPR));
if (hasMAIInsts) {
- EmitResourceUsageRemark(
- "NumAGPR", "AGPRs",
- getMCExprValue(CurrentProgramInfo.NumAccVGPR, MCCtx));
+ EmitResourceUsageRemark("NumAGPR", "AGPRs",
+ getMCExprStr(CurrentProgramInfo.NumAccVGPR));
}
- EmitResourceUsageRemark(
- "ScratchSize", "ScratchSize [bytes/lane]",
- getMCExprValue(CurrentProgramInfo.ScratchSize, MCCtx));
+ EmitResourceUsageRemark("ScratchSize", "ScratchSize [bytes/lane]",
+ getMCExprStr(CurrentProgramInfo.ScratchSize));
+ int64_t DynStack;
+ bool DynStackEvaluatable =
+ CurrentProgramInfo.DynamicCallStack->evaluateAsAbsolute(DynStack);
StringRef DynamicStackStr =
- getMCExprValue(CurrentProgramInfo.DynamicCallStack, MCCtx) ? "True"
- : "False";
+ DynStackEvaluatable && DynStack ? "True" : "False";
EmitResourceUsageRemark("DynamicStack", "Dynamic Stack", DynamicStackStr);
EmitResourceUsageRemark("Occupancy", "Occupancy [waves/SIMD]",
- getMCExprValue(CurrentProgramInfo.Occupancy, MCCtx));
+ getMCExprStr(CurrentProgramInfo.Occupancy));
EmitResourceUsageRemark("SGPRSpill", "SGPRs Spill",
CurrentProgramInfo.SGPRSpill);
EmitResourceUsageRemark("VGPRSpill", "VGPRs Spill",
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.h b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.h
index 87156f27fc6c5..2a3a39029dd2a 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.h
+++ b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.h
@@ -65,6 +65,12 @@ class AMDGPUAsmPrinter final : public AsmPrinter {
uint32_t TotalNumVGPR, uint32_t NumSGPR,
uint64_t ScratchSize, uint64_t CodeSize,
const AMDGPUMachineFunction *MFI);
+ void emitCommonFunctionComments(const MCExpr *NumVGPR,
+ std::optional<const MCExpr *> NumAGPR,
+ const MCExpr *TotalNumVGPR,
+ const MCExpr *NumSGPR,
+ const MCExpr *ScratchSize, uint64_t CodeSize,
+ const AMDGPUMachineFunction *MFI);
void emitResourceUsageRemarks(const MachineFunction &MF,
const SIProgramInfo &CurrentProgramInfo,
bool isModuleEntryFunction, bool hasMAIInsts);
@@ -79,6 +85,7 @@ class AMDGPUAsmPrinter final : public AsmPrinter {
void initTargetStreamer(Module &M);
static uint64_t getMCExprValue(const MCExpr *Value, MCContext &Ctx);
+ std::string getMCExprStr(const MCExpr *Value);
public:
explicit AMDGPUAsmPrinter(TargetMachine &TM,
|
arsenm
reviewed
Jun 6, 2024
arsenm
reviewed
Jun 7, 2024
| std::string AMDGPUAsmPrinter::getMCExprStr(const MCExpr *Value) { | ||
| std::string Str; | ||
| raw_string_ostream OSS(Str); | ||
| StringRef AMDGPUAsmPrinter::getMCExprStr(const MCExpr *Value) { |
There was a problem hiding this comment.
You can't return this StringRef references to the local SmallString. You have to return the SmallString
arsenm
approved these changes
Jun 7, 2024
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Eliminates assumption that MCExpr comments/remarks being emitted are always resolvable