[IRBuilder] Add new overload for CreateIntrinsic #131942
Open
Conversation
This file contains 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
jurahul
force-pushed
the
new_createintrinsic_overload
branch
from
caacc1f to
e24106e
Compare
llvmbot
added
clang
|
@llvm/pr-subscribers-hlsl @llvm/pr-subscribers-clang Author: Rahul Joshi (jurahul) ChangesAdd a new Patch is 58.77 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/131942.diff 41 Files Affected:
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index c126f88b9e3a5..8b11175f7c5d2 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -114,13 +114,14 @@ static Value *handleHlslClip(const CallExpr *E, CodeGenFunction *CGF) {
CMP = CGF->Builder.CreateIntrinsic(
CGF->Builder.getInt1Ty(), CGF->CGM.getHLSLRuntime().getAnyIntrinsic(),
{FCompInst}, nullptr);
- } else
+ } else {
CMP = CGF->Builder.CreateFCmpOLT(Op0, FZeroConst);
+ }
- if (CGF->CGM.getTarget().getTriple().isDXIL())
+ if (CGF->CGM.getTarget().getTriple().isDXIL()) {
LastInstr = CGF->Builder.CreateIntrinsic(
CGF->VoidTy, llvm::Intrinsic::dx_discard, {CMP}, nullptr);
- else if (CGF->CGM.getTarget().getTriple().isSPIRV()) {
+ } else if (CGF->CGM.getTarget().getTriple().isSPIRV()) {
BasicBlock *LT0 = CGF->createBasicBlock("lt0", CGF->CurFn);
BasicBlock *End = CGF->createBasicBlock("end", CGF->CurFn);
@@ -128,8 +129,8 @@ static Value *handleHlslClip(const CallExpr *E, CodeGenFunction *CGF) {
CGF->Builder.SetInsertPoint(LT0);
- CGF->Builder.CreateIntrinsic(CGF->VoidTy, llvm::Intrinsic::spv_discard, {},
- nullptr);
+ CGF->Builder.CreateIntrinsic(llvm::Intrinsic::spv_discard,
+ ArrayRef<Value *>{}, nullptr);
LastInstr = CGF->Builder.CreateBr(End);
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.cpp b/clang/lib/CodeGen/CGHLSLRuntime.cpp
index a273f1e50c8b5..00f78c9ded116 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.cpp
+++ b/clang/lib/CodeGen/CGHLSLRuntime.cpp
@@ -386,8 +386,8 @@ void CGHLSLRuntime::emitEntryFunction(const FunctionDecl *FD,
SmallVector<OperandBundleDef, 1> OB;
if (CGM.shouldEmitConvergenceTokens()) {
assert(EntryFn->isConvergent());
- llvm::Value *I = B.CreateIntrinsic(
- llvm::Intrinsic::experimental_convergence_entry, {}, {});
+ llvm::Value *I =
+ B.CreateIntrinsic(llvm::Intrinsic::experimental_convergence_entry, {});
llvm::Value *bundleArgs[] = {I};
OB.emplace_back("convergencectrl", bundleArgs);
}
diff --git a/llvm/include/llvm/IR/IRBuilder.h b/llvm/include/llvm/IR/IRBuilder.h
index 750a99cc50dd7..07660e93253da 100644
--- a/llvm/include/llvm/IR/IRBuilder.h
+++ b/llvm/include/llvm/IR/IRBuilder.h
@@ -1004,6 +1004,14 @@ class IRBuilderBase {
ArrayRef<Value *> Args, FMFSource FMFSource = {},
const Twine &Name = "");
+ /// Create a call to non-overloaded intrinsic \p ID with \p Args. If
+ /// \p FMFSource is provided, copy fast-math-flags from that instruction to
+ /// the intrinsic.
+ CallInst *CreateIntrinsic(Intrinsic::ID ID, ArrayRef<Value *> Args,
+ FMFSource FMFSource = {}, const Twine &Name = "") {
+ return CreateIntrinsic(ID, /*Types=*/{}, Args, FMFSource, Name);
+ }
+
/// Create call to the minnum intrinsic.
Value *CreateMinNum(Value *LHS, Value *RHS, FMFSource FMFSource = {},
const Twine &Name = "") {
diff --git a/llvm/lib/CodeGen/SafeStack.cpp b/llvm/lib/CodeGen/SafeStack.cpp
index 74e9d945c1885..da229f86f24ce 100644
--- a/llvm/lib/CodeGen/SafeStack.cpp
+++ b/llvm/lib/CodeGen/SafeStack.cpp
@@ -367,7 +367,7 @@ Value *SafeStack::getStackGuard(IRBuilder<> &IRB, Function &F) {
if (!StackGuardVar) {
TL.insertSSPDeclarations(*M);
- return IRB.CreateIntrinsic(Intrinsic::stackguard, {}, {});
+ return IRB.CreateIntrinsic(Intrinsic::stackguard, {});
}
return IRB.CreateLoad(StackPtrTy, StackGuardVar, "StackGuard");
diff --git a/llvm/lib/CodeGen/StackProtector.cpp b/llvm/lib/CodeGen/StackProtector.cpp
index e823df3186a54..4c8b21d1e29aa 100644
--- a/llvm/lib/CodeGen/StackProtector.cpp
+++ b/llvm/lib/CodeGen/StackProtector.cpp
@@ -542,7 +542,7 @@ static Value *getStackGuard(const TargetLoweringBase *TLI, Module *M,
if (SupportsSelectionDAGSP)
*SupportsSelectionDAGSP = true;
TLI->insertSSPDeclarations(*M);
- return B.CreateIntrinsic(Intrinsic::stackguard, {}, {});
+ return B.CreateIntrinsic(Intrinsic::stackguard, {});
}
/// Insert code into the entry block that stores the stack guard
@@ -563,7 +563,7 @@ static bool CreatePrologue(Function *F, Module *M, Instruction *CheckLoc,
AI = B.CreateAlloca(PtrTy, nullptr, "StackGuardSlot");
Value *GuardSlot = getStackGuard(TLI, M, B, &SupportsSelectionDAGSP);
- B.CreateIntrinsic(Intrinsic::stackprotector, {}, {GuardSlot, AI});
+ B.CreateIntrinsic(Intrinsic::stackprotector, {GuardSlot, AI});
return SupportsSelectionDAGSP;
}
diff --git a/llvm/lib/IR/AutoUpgrade.cpp b/llvm/lib/IR/AutoUpgrade.cpp
index 7a194219c5cd4..d0e043b1e582e 100644
--- a/llvm/lib/IR/AutoUpgrade.cpp
+++ b/llvm/lib/IR/AutoUpgrade.cpp
@@ -1766,7 +1766,7 @@ static Value *upgradeX86VPERMT2Intrinsics(IRBuilder<> &Builder, CallBase &CI,
if (!IndexForm)
std::swap(Args[0], Args[1]);
- Value *V = Builder.CreateIntrinsic(IID, {}, Args);
+ Value *V = Builder.CreateIntrinsic(IID, Args);
Value *PassThru = ZeroMask ? ConstantAggregateZero::get(Ty)
: Builder.CreateBitCast(CI.getArgOperand(1),
Ty);
@@ -2021,8 +2021,8 @@ static Value *upgradeMaskedCompare(IRBuilder<> &Builder, CallBase &CI,
// Replace a masked intrinsic with an older unmasked intrinsic.
static Value *upgradeX86MaskedShift(IRBuilder<> &Builder, CallBase &CI,
Intrinsic::ID IID) {
- Value *Rep = Builder.CreateIntrinsic(
- IID, {}, {CI.getArgOperand(0), CI.getArgOperand(1)});
+ Value *Rep =
+ Builder.CreateIntrinsic(IID, {CI.getArgOperand(0), CI.getArgOperand(1)});
return emitX86Select(Builder, CI.getArgOperand(3), Rep, CI.getArgOperand(2));
}
@@ -2279,7 +2279,7 @@ static bool upgradeAVX512MaskToSelect(StringRef Name, IRBuilder<> &Builder,
SmallVector<Value *, 4> Args(CI.args());
Args.pop_back();
Args.pop_back();
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
unsigned NumArgs = CI.arg_size();
Rep = emitX86Select(Builder, CI.getArgOperand(NumArgs - 1), Rep,
CI.getArgOperand(NumArgs - 2));
@@ -2504,7 +2504,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
: Intrinsic::x86_avx512_sqrt_pd_512;
Value *Args[] = {CI->getArgOperand(0), CI->getArgOperand(3)};
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
} else {
Rep = Builder.CreateIntrinsic(Intrinsic::sqrt, CI->getType(),
{CI->getArgOperand(0)});
@@ -2631,8 +2631,8 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
break;
}
- Rep = Builder.CreateIntrinsic(IID, {},
- {CI->getOperand(0), CI->getArgOperand(1)});
+ Rep =
+ Builder.CreateIntrinsic(IID, {CI->getOperand(0), CI->getArgOperand(1)});
Rep = applyX86MaskOn1BitsVec(Builder, Rep, CI->getArgOperand(2));
} else if (Name.starts_with("avx512.mask.fpclass.p")) {
Type *OpTy = CI->getArgOperand(0)->getType();
@@ -2654,8 +2654,8 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
else
llvm_unreachable("Unexpected intrinsic");
- Rep = Builder.CreateIntrinsic(IID, {},
- {CI->getOperand(0), CI->getArgOperand(1)});
+ Rep =
+ Builder.CreateIntrinsic(IID, {CI->getOperand(0), CI->getArgOperand(1)});
Rep = applyX86MaskOn1BitsVec(Builder, Rep, CI->getArgOperand(2));
} else if (Name.starts_with("avx512.cmp.p")) {
SmallVector<Value *, 4> Args(CI->args());
@@ -2683,7 +2683,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
std::swap(Mask, Args.back());
Args.push_back(Mask);
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
} else if (Name.starts_with("avx512.mask.cmp.")) {
// Integer compare intrinsics.
unsigned Imm = cast<ConstantInt>(CI->getArgOperand(2))->getZExtValue();
@@ -2899,7 +2899,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
} else if (Name == "sse42.crc32.64.8") {
Value *Trunc0 =
Builder.CreateTrunc(CI->getArgOperand(0), Type::getInt32Ty(C));
- Rep = Builder.CreateIntrinsic(Intrinsic::x86_sse42_crc32_32_8, {},
+ Rep = Builder.CreateIntrinsic(Intrinsic::x86_sse42_crc32_32_8,
{Trunc0, CI->getArgOperand(1)});
Rep = Builder.CreateZExt(Rep, CI->getType(), "");
} else if (Name.starts_with("avx.vbroadcast.s") ||
@@ -3389,7 +3389,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
IID = Intrinsic::x86_avx512_add_pd_512;
Rep = Builder.CreateIntrinsic(
- IID, {},
+ IID,
{CI->getArgOperand(0), CI->getArgOperand(1), CI->getArgOperand(4)});
} else {
Rep = Builder.CreateFAdd(CI->getArgOperand(0), CI->getArgOperand(1));
@@ -3405,7 +3405,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
IID = Intrinsic::x86_avx512_div_pd_512;
Rep = Builder.CreateIntrinsic(
- IID, {},
+ IID,
{CI->getArgOperand(0), CI->getArgOperand(1), CI->getArgOperand(4)});
} else {
Rep = Builder.CreateFDiv(CI->getArgOperand(0), CI->getArgOperand(1));
@@ -3421,7 +3421,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
IID = Intrinsic::x86_avx512_mul_pd_512;
Rep = Builder.CreateIntrinsic(
- IID, {},
+ IID,
{CI->getArgOperand(0), CI->getArgOperand(1), CI->getArgOperand(4)});
} else {
Rep = Builder.CreateFMul(CI->getArgOperand(0), CI->getArgOperand(1));
@@ -3437,7 +3437,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
IID = Intrinsic::x86_avx512_sub_pd_512;
Rep = Builder.CreateIntrinsic(
- IID, {},
+ IID,
{CI->getArgOperand(0), CI->getArgOperand(1), CI->getArgOperand(4)});
} else {
Rep = Builder.CreateFSub(CI->getArgOperand(0), CI->getArgOperand(1));
@@ -3455,7 +3455,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
Intrinsic::ID IID = MinMaxTbl[IsMin][IsDouble];
Rep = Builder.CreateIntrinsic(
- IID, {},
+ IID,
{CI->getArgOperand(0), CI->getArgOperand(1), CI->getArgOperand(4)});
Rep =
emitX86Select(Builder, CI->getArgOperand(3), Rep, CI->getArgOperand(2));
@@ -3753,7 +3753,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
IID = Intrinsic::x86_avx512_vfmadd_f64;
else
IID = Intrinsic::x86_avx512_vfmadd_f32;
- Rep = Builder.CreateIntrinsic(IID, {}, Ops);
+ Rep = Builder.CreateIntrinsic(IID, Ops);
} else {
Rep = Builder.CreateFMA(A, B, C);
}
@@ -3806,7 +3806,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
else
IID = Intrinsic::x86_avx512_vfmadd_pd_512;
- Rep = Builder.CreateIntrinsic(IID, {}, {A, B, C, CI->getArgOperand(4)});
+ Rep = Builder.CreateIntrinsic(IID, {A, B, C, CI->getArgOperand(4)});
} else {
Rep = Builder.CreateFMA(A, B, C);
}
@@ -3834,7 +3834,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
Value *Ops[] = {CI->getArgOperand(0), CI->getArgOperand(1),
CI->getArgOperand(2)};
Ops[2] = Builder.CreateFNeg(Ops[2]);
- Rep = Builder.CreateIntrinsic(IID, {}, Ops);
+ Rep = Builder.CreateIntrinsic(IID, Ops);
} else if (Name.starts_with("avx512.mask.vfmaddsub.p") ||
Name.starts_with("avx512.mask3.vfmaddsub.p") ||
Name.starts_with("avx512.maskz.vfmaddsub.p") ||
@@ -3857,7 +3857,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
if (IsSubAdd)
Ops[2] = Builder.CreateFNeg(Ops[2]);
- Rep = Builder.CreateIntrinsic(IID, {}, Ops);
+ Rep = Builder.CreateIntrinsic(IID, Ops);
} else {
int NumElts = cast<FixedVectorType>(CI->getType())->getNumElements();
@@ -3908,7 +3908,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
Value *Args[] = {CI->getArgOperand(0), CI->getArgOperand(1),
CI->getArgOperand(2), CI->getArgOperand(3)};
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
Value *PassThru = ZeroMask ? ConstantAggregateZero::get(CI->getType())
: CI->getArgOperand(0);
Rep = emitX86Select(Builder, CI->getArgOperand(4), Rep, PassThru);
@@ -3935,7 +3935,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
Value *Args[] = {CI->getArgOperand(0), CI->getArgOperand(1),
CI->getArgOperand(2)};
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
Value *PassThru = ZeroMask ? ConstantAggregateZero::get(CI->getType())
: CI->getArgOperand(0);
Rep = emitX86Select(Builder, CI->getArgOperand(3), Rep, PassThru);
@@ -3970,7 +3970,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
Value *Args[] = {CI->getArgOperand(0), CI->getArgOperand(1),
CI->getArgOperand(2)};
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
Value *PassThru = ZeroMask ? ConstantAggregateZero::get(CI->getType())
: CI->getArgOperand(0);
Rep = emitX86Select(Builder, CI->getArgOperand(3), Rep, PassThru);
@@ -3999,7 +3999,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
Value *Args[] = {CI->getArgOperand(0), CI->getArgOperand(1),
CI->getArgOperand(2)};
- Rep = Builder.CreateIntrinsic(IID, {}, Args);
+ Rep = Builder.CreateIntrinsic(IID, Args);
Value *PassThru = ZeroMask ? ConstantAggregateZero::get(CI->getType())
: CI->getArgOperand(0);
Rep = emitX86Select(Builder, CI->getArgOperand(3), Rep, PassThru);
@@ -4021,7 +4021,7 @@ static Value *upgradeX86IntrinsicCall(StringRef Name, CallBase *CI, Function *F,
// Make a call with 3 operands.
Value *Args[] = {CI->getArgOperand(0), CI->getArgOperand(1),
CI->getArgOperand(2)};
- Value *NewCall = Builder.CreateIntrinsic(IID, {}, Args);
+ Value *NewCall = Builder.CreateIntrinsic(IID, Args);
// Extract the second result and store it.
Value *Data = Builder.CreateExtractValue(NewCall, 1);
@@ -4089,7 +4089,7 @@ static Value *upgradeAArch64IntrinsicCall(StringRef Name, CallBase *CI,
Args[1] = Builder.CreateIntrinsic(
Intrinsic::aarch64_sve_convert_from_svbool, GoodPredTy, Args[1]);
- return Builder.CreateIntrinsic(NewID, {}, Args, /*FMFSource=*/nullptr,
+ return Builder.CreateIntrinsic(NewID, Args, /*FMFSource=*/nullptr,
CI->getName());
}
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index 879b83f94b79a..a1e260bc81aed 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -28059,7 +28059,7 @@ Value *AArch64TargetLowering::emitLoadLinked(IRBuilderBase &Builder,
IsAcquire ? Intrinsic::aarch64_ldaxp : Intrinsic::aarch64_ldxp;
Value *LoHi =
- Builder.CreateIntrinsic(Int, {}, Addr, /*FMFSource=*/nullptr, "lohi");
+ Builder.CreateIntrinsic(Int, Addr, /*FMFSource=*/nullptr, "lohi");
Value *Lo = Builder.CreateExtractValue(LoHi, 0, "lo");
Value *Hi = Builder.CreateExtractValue(LoHi, 1, "hi");
@@ -28089,7 +28089,7 @@ Value *AArch64TargetLowering::emitLoadLinked(IRBuilderBase &Builder,
void AArch64TargetLowering::emitAtomicCmpXchgNoStoreLLBalance(
IRBuilderBase &Builder) const {
- Builder.CreateIntrinsic(Intrinsic::aarch64_clrex, {}, {});
+ Builder.CreateIntrinsic(Intrinsic::aarch64_clrex, {});
}
Value *AArch64TargetLowering::emitStoreConditional(IRBuilderBase &Builder,
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index 7cec8a17dfaaa..967c5abeeb4ee 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -1476,7 +1476,7 @@ static std::optional<Instruction *> instCombineRDFFR(InstCombiner &IC,
auto *PTrue = IC.Builder.CreateIntrinsic(Intrinsic::aarch64_sve_ptrue,
{II.getType()}, {AllPat});
auto *RDFFR =
- IC.Builder.CreateIntrinsic(Intrinsic::aarch64_sve_rdffr_z, {}, {PTrue});
+ IC.Builder.CreateIntrinsic(Intrinsic::aarch64_sve_rdffr_z, {PTrue});
RDFFR->takeName(&II);
return IC.replaceInstUsesWith(II, RDFFR);
}
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAsanInstrumentation.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAsanInstrumentation.cpp
index 6554863e08c91..19e2a6a27020d 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUAsanInstrumentation.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUAsanInstrumentation.cpp
@@ -73,7 +73,7 @@ static Instruction *genAMDGPUReportBlock(Module &M, IRBuilder<> &IRB,
Trm = SplitBlockAndInsertIfThen(Cond, Trm, false);
IRB.SetInsertPoint(Trm);
- return IRB.CreateIntrinsic(Intrinsic::amdgcn_unreachable, {}, {});
+ return IRB.CreateIntrinsic(Intrinsic::amdgcn_unreachable, {});
}
static Value *createSlowPathCmp(Module &M, IRBuilder<> &IRB, Type *IntptrTy,
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp
index 76b1775f0d096..0a163f8dc7f6b 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp
@@ -666,7 +666,7 @@ void AMDGPUAtomicOptimizerImpl::optimizeAtomic(Instruction &I,
// Record I's original position as the entry block.
PixelEntryBB = I.getParent();
- Value *const Cond = B.CreateIntrinsic(Intrinsic::amdgcn_ps_live, {}, {});
+ Value *const Cond = B.CreateIntrinsic(Intrinsic::amdgcn_ps_live, {});
Instruction *const NonHelperTerminator =
SplitBlockAndInsertIfThen(Cond, &I, false, nullptr, &DTU, nullptr);
@@ -698,15 +698,14 @@ void AMDGPUAtomicOptimizerImpl::optimizeAtomic(Instruction &I,
// using the mbcnt intrinsic.
Value *Mbcnt;
if (ST.isWave32()) {
- Mbcnt = B.CreateIntrinsic(Intrinsic::amdgcn_mbcnt_lo, {},
- {Ballot, B.getInt32(0)});
+ Mbcnt =
+ B.CreateIntrinsic(Intrinsic::amdgcn_mbcnt_lo, {Ballot, B.getInt32(0)});
} else {
Value *const ExtractLo = B.CreateTrunc(Ballot, Int32Ty);
Value *const ExtractHi = B.CreateTrunc(B.CreateLShr(Ballot, 32), Int32Ty);
- Mbcnt = B.CreateIntrinsic(Intrinsic::amdgcn_mbcnt_lo, {},
+ Mbcnt = B.CreateIntrinsic(Intrinsic::amdgcn_mbcnt_lo,
{ExtractLo, B.getInt32(0)});
- Mbcnt =
- B.CreateIntrinsic(Intrinsic::amdgcn_mbcnt_hi, {}, {ExtractHi, Mbcnt});
+ Mbcnt = B.CreateIntrinsic(Intrinsic::amdgcn_mbcnt_hi, {ExtractHi, Mbcnt});
}
Function *F = I.getFunction();
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp b/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
index fdba8835cbf0a..a0c28879f6df4 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
@@ -1033,7 +1033,7 @@ Value *AMDGPUCodeGenPrepareImpl::optimizeWithFDivFast(
if (!HasFP32DenormalFlush && !NumIsOne)
return nullptr;
- return Builder.CreateIntrinsic(Intrinsic::amdgcn_fdiv_fast, {}, {Num, Den});
+ return Builder.CreateIntrinsic(Intrinsic::amdgcn_fdiv_fast, {Num, Den});
}
Value *AMDGPUCodeGenPrepareImpl::visitFDivElement(
diff --git a/llvm/lib/Target/AMDGPU/AMDGPULowerKernelArguments.cpp b/llvm/lib/Target/AMDGPU/AMDGPULowerKernelArguments.cpp
index 09412d1b0f1cc..a4e6768b4...
[truncated]
|
jayfoad
approved these changes
Mar 20, 2025
jurahul
force-pushed
the
new_createintrinsic_overload
branch
from
e24106e to
7f7f0ea
Compare
Add a new `CreateIntrinsic` overload with no `Types`, useful for creating calls to non-overloaded intrinsics that don't need additional mangling.
jurahul
force-pushed
the
new_createintrinsic_overload
branch
from
7f7f0ea to
321ac98
Compare
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.
Add a new
CreateIntrinsicoverload with noTypes, useful for creating calls to non-overloaded intrinsics that don't need additional mangling.