Clang/buildFMulAdd: Use negated attribute #121038
Draft
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
This reverts commit 67789aa.
Use negated attribute if negMul or negAdd. So that we can lower fneg+fmuladd to fmul+fsub if needed. 1) It can save one machine instruction: fneg/fmul/fadd vs fmul/fsub 2) In strict mode, `c-a*b` may be different with `c+(-a)*b`.
llvmbot added clang 
Member
| @llvm/pr-subscribers-llvm-ir @llvm/pr-subscribers-clang-codegen Author: YunQiang Su (wzssyqa) ChangesUse negated attribute if negMul or negAdd. So that we can lower
Full diff: https://github.com/llvm/llvm-project/pull/121038.diff 7 Files Affected:
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp index 4b71bd730ce12c..14d73de055d8ec 100644 --- a/clang/lib/CodeGen/CGExprScalar.cpp +++ b/clang/lib/CodeGen/CGExprScalar.cpp @@ -4120,6 +4120,10 @@ static Value* buildFMulAdd(llvm::Instruction *MulOp, Value *Addend, CGF.CGM.getIntrinsic(llvm::Intrinsic::experimental_constrained_fmuladd, Addend->getType()), {MulOp0, MulOp1, Addend}); + if (negMul) + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(0, llvm::Attribute::Negated); + if (negAdd) + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(2, llvm::Attribute::Negated); } else { FMulAdd = Builder.CreateCall( CGF.CGM.getIntrinsic(llvm::Intrinsic::fmuladd, Addend->getType()), diff --git a/clang/test/CodeGen/constrained-math-builtins.c b/clang/test/CodeGen/constrained-math-builtins.c index 68b9e75283c547..f044f15e98918b 100644 --- a/clang/test/CodeGen/constrained-math-builtins.c +++ b/clang/test/CodeGen/constrained-math-builtins.c @@ -392,12 +392,12 @@ void bar(float f) { // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg - // CHECK: call double @llvm.experimental.constrained.fmuladd.f64(double %{{.*}}, double %{{.*}}, double %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call double @llvm.experimental.constrained.fmuladd.f64(double %{{.*}}, double %{{.*}}, double negated %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg // CHECK: call x86_fp80 @llvm.experimental.constrained.fmuladd.f80(x86_fp80 %{{.*}}, x86_fp80 %{{.*}}, x86_fp80 %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg // CHECK: fneg - // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float negated %{{.*}}, float %{{.*}}, float negated %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg - // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float negated %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") }; diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst index 7e01331b20c570..bf37e6a788c4b6 100644 --- a/llvm/docs/LangRef.rst +++ b/llvm/docs/LangRef.rst @@ -1573,6 +1573,12 @@ Currently, only the following parameter attributes are defined: | pinf | Positive infinity | 512 | +-------+----------------------+---------------+ +``negated`` + The function parameter marked with this attribute is negated from + its opposite number by the frontend like Clang. The middle end or + backend should convert it back if possible. For example if -(a*b) + is converted to (-a)*b, the arg0 of `fmul` instruction should be + marked with `negated` attribute. ``alignstack(<n>)`` This indicates the alignment that should be considered by the backend when diff --git a/llvm/include/llvm/Bitcode/LLVMBitCodes.h b/llvm/include/llvm/Bitcode/LLVMBitCodes.h index 21fd27d9838db7..7e9d174db22026 100644 --- a/llvm/include/llvm/Bitcode/LLVMBitCodes.h +++ b/llvm/include/llvm/Bitcode/LLVMBitCodes.h @@ -788,6 +788,7 @@ enum AttributeKindCodes { ATTR_KIND_NO_EXT = 99, ATTR_KIND_NO_DIVERGENCE_SOURCE = 100, ATTR_KIND_SANITIZE_TYPE = 101, + ATTR_KIND_NEGATED = 102, }; enum ComdatSelectionKindCodes { diff --git a/llvm/include/llvm/IR/Attributes.td b/llvm/include/llvm/IR/Attributes.td index 61955cf883c3f1..baeca5d53f3c46 100644 --- a/llvm/include/llvm/IR/Attributes.td +++ b/llvm/include/llvm/IR/Attributes.td @@ -162,6 +162,9 @@ def Memory : IntAttr<"memory", IntersectCustom, [FnAttr]>; /// Forbidden floating-point classes. def NoFPClass : IntAttr<"nofpclass", IntersectCustom, [ParamAttr, RetAttr]>; +/// Converted from the opposite number +def Negated : EnumAttr<"negated", IntersectAnd, [ParamAttr, RetAttr]>; + /// Function must be optimized for size first. def MinSize : EnumAttr<"minsize", IntersectPreserve, [FnAttr]>; diff --git a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp index b4efd3928a2e6f..e87c9d2e13883d 100644 --- a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp +++ b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp @@ -755,6 +755,8 @@ static uint64_t getAttrKindEncoding(Attribute::AttrKind Kind) { return bitc::ATTR_KIND_MEMORY; case Attribute::NoFPClass: return bitc::ATTR_KIND_NOFPCLASS; + case Attribute::Negated: + return bitc::ATTR_KIND_NEGATED; case Attribute::Naked: return bitc::ATTR_KIND_NAKED; case Attribute::Nest: diff --git a/llvm/lib/Transforms/Utils/CodeExtractor.cpp b/llvm/lib/Transforms/Utils/CodeExtractor.cpp index 7ddb9e22c83441..4e1a8c560078aa 100644 --- a/llvm/lib/Transforms/Utils/CodeExtractor.cpp +++ b/llvm/lib/Transforms/Utils/CodeExtractor.cpp @@ -918,6 +918,7 @@ Function *CodeExtractor::constructFunctionDeclaration( case Attribute::PresplitCoroutine: case Attribute::Memory: case Attribute::NoFPClass: + case Attribute::Negated: case Attribute::CoroDestroyOnlyWhenComplete: case Attribute::CoroElideSafe: case Attribute::NoDivergenceSource: |
Member
| @llvm/pr-subscribers-clang Author: YunQiang Su (wzssyqa) ChangesUse negated attribute if negMul or negAdd. So that we can lower
Full diff: https://github.com/llvm/llvm-project/pull/121038.diff 7 Files Affected:
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp index 4b71bd730ce12c..14d73de055d8ec 100644 --- a/clang/lib/CodeGen/CGExprScalar.cpp +++ b/clang/lib/CodeGen/CGExprScalar.cpp @@ -4120,6 +4120,10 @@ static Value* buildFMulAdd(llvm::Instruction *MulOp, Value *Addend, CGF.CGM.getIntrinsic(llvm::Intrinsic::experimental_constrained_fmuladd, Addend->getType()), {MulOp0, MulOp1, Addend}); + if (negMul) + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(0, llvm::Attribute::Negated); + if (negAdd) + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(2, llvm::Attribute::Negated); } else { FMulAdd = Builder.CreateCall( CGF.CGM.getIntrinsic(llvm::Intrinsic::fmuladd, Addend->getType()), diff --git a/clang/test/CodeGen/constrained-math-builtins.c b/clang/test/CodeGen/constrained-math-builtins.c index 68b9e75283c547..f044f15e98918b 100644 --- a/clang/test/CodeGen/constrained-math-builtins.c +++ b/clang/test/CodeGen/constrained-math-builtins.c @@ -392,12 +392,12 @@ void bar(float f) { // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg - // CHECK: call double @llvm.experimental.constrained.fmuladd.f64(double %{{.*}}, double %{{.*}}, double %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call double @llvm.experimental.constrained.fmuladd.f64(double %{{.*}}, double %{{.*}}, double negated %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg // CHECK: call x86_fp80 @llvm.experimental.constrained.fmuladd.f80(x86_fp80 %{{.*}}, x86_fp80 %{{.*}}, x86_fp80 %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg // CHECK: fneg - // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float negated %{{.*}}, float %{{.*}}, float negated %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg - // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float negated %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") }; diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst index 7e01331b20c570..bf37e6a788c4b6 100644 --- a/llvm/docs/LangRef.rst +++ b/llvm/docs/LangRef.rst @@ -1573,6 +1573,12 @@ Currently, only the following parameter attributes are defined: | pinf | Positive infinity | 512 | +-------+----------------------+---------------+ +``negated`` + The function parameter marked with this attribute is negated from + its opposite number by the frontend like Clang. The middle end or + backend should convert it back if possible. For example if -(a*b) + is converted to (-a)*b, the arg0 of `fmul` instruction should be + marked with `negated` attribute. ``alignstack(<n>)`` This indicates the alignment that should be considered by the backend when diff --git a/llvm/include/llvm/Bitcode/LLVMBitCodes.h b/llvm/include/llvm/Bitcode/LLVMBitCodes.h index 21fd27d9838db7..7e9d174db22026 100644 --- a/llvm/include/llvm/Bitcode/LLVMBitCodes.h +++ b/llvm/include/llvm/Bitcode/LLVMBitCodes.h @@ -788,6 +788,7 @@ enum AttributeKindCodes { ATTR_KIND_NO_EXT = 99, ATTR_KIND_NO_DIVERGENCE_SOURCE = 100, ATTR_KIND_SANITIZE_TYPE = 101, + ATTR_KIND_NEGATED = 102, }; enum ComdatSelectionKindCodes { diff --git a/llvm/include/llvm/IR/Attributes.td b/llvm/include/llvm/IR/Attributes.td index 61955cf883c3f1..baeca5d53f3c46 100644 --- a/llvm/include/llvm/IR/Attributes.td +++ b/llvm/include/llvm/IR/Attributes.td @@ -162,6 +162,9 @@ def Memory : IntAttr<"memory", IntersectCustom, [FnAttr]>; /// Forbidden floating-point classes. def NoFPClass : IntAttr<"nofpclass", IntersectCustom, [ParamAttr, RetAttr]>; +/// Converted from the opposite number +def Negated : EnumAttr<"negated", IntersectAnd, [ParamAttr, RetAttr]>; + /// Function must be optimized for size first. def MinSize : EnumAttr<"minsize", IntersectPreserve, [FnAttr]>; diff --git a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp index b4efd3928a2e6f..e87c9d2e13883d 100644 --- a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp +++ b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp @@ -755,6 +755,8 @@ static uint64_t getAttrKindEncoding(Attribute::AttrKind Kind) { return bitc::ATTR_KIND_MEMORY; case Attribute::NoFPClass: return bitc::ATTR_KIND_NOFPCLASS; + case Attribute::Negated: + return bitc::ATTR_KIND_NEGATED; case Attribute::Naked: return bitc::ATTR_KIND_NAKED; case Attribute::Nest: diff --git a/llvm/lib/Transforms/Utils/CodeExtractor.cpp b/llvm/lib/Transforms/Utils/CodeExtractor.cpp index 7ddb9e22c83441..4e1a8c560078aa 100644 --- a/llvm/lib/Transforms/Utils/CodeExtractor.cpp +++ b/llvm/lib/Transforms/Utils/CodeExtractor.cpp @@ -918,6 +918,7 @@ Function *CodeExtractor::constructFunctionDeclaration( case Attribute::PresplitCoroutine: case Attribute::Memory: case Attribute::NoFPClass: + case Attribute::Negated: case Attribute::CoroDestroyOnlyWhenComplete: case Attribute::CoroElideSafe: case Attribute::NoDivergenceSource: |
Member
| @llvm/pr-subscribers-llvm-transforms Author: YunQiang Su (wzssyqa) ChangesUse negated attribute if negMul or negAdd. So that we can lower
Full diff: https://github.com/llvm/llvm-project/pull/121038.diff 7 Files Affected:
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp index 4b71bd730ce12c..14d73de055d8ec 100644 --- a/clang/lib/CodeGen/CGExprScalar.cpp +++ b/clang/lib/CodeGen/CGExprScalar.cpp @@ -4120,6 +4120,10 @@ static Value* buildFMulAdd(llvm::Instruction *MulOp, Value *Addend, CGF.CGM.getIntrinsic(llvm::Intrinsic::experimental_constrained_fmuladd, Addend->getType()), {MulOp0, MulOp1, Addend}); + if (negMul) + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(0, llvm::Attribute::Negated); + if (negAdd) + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(2, llvm::Attribute::Negated); } else { FMulAdd = Builder.CreateCall( CGF.CGM.getIntrinsic(llvm::Intrinsic::fmuladd, Addend->getType()), diff --git a/clang/test/CodeGen/constrained-math-builtins.c b/clang/test/CodeGen/constrained-math-builtins.c index 68b9e75283c547..f044f15e98918b 100644 --- a/clang/test/CodeGen/constrained-math-builtins.c +++ b/clang/test/CodeGen/constrained-math-builtins.c @@ -392,12 +392,12 @@ void bar(float f) { // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg - // CHECK: call double @llvm.experimental.constrained.fmuladd.f64(double %{{.*}}, double %{{.*}}, double %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call double @llvm.experimental.constrained.fmuladd.f64(double %{{.*}}, double %{{.*}}, double negated %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg // CHECK: call x86_fp80 @llvm.experimental.constrained.fmuladd.f80(x86_fp80 %{{.*}}, x86_fp80 %{{.*}}, x86_fp80 %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg // CHECK: fneg - // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float negated %{{.*}}, float %{{.*}}, float negated %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") // CHECK: fneg - // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") + // CHECK: call float @llvm.experimental.constrained.fmuladd.f32(float negated %{{.*}}, float %{{.*}}, float %{{.*}}, metadata !"round.tonearest", metadata !"fpexcept.strict") }; diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst index 7e01331b20c570..bf37e6a788c4b6 100644 --- a/llvm/docs/LangRef.rst +++ b/llvm/docs/LangRef.rst @@ -1573,6 +1573,12 @@ Currently, only the following parameter attributes are defined: | pinf | Positive infinity | 512 | +-------+----------------------+---------------+ +``negated`` + The function parameter marked with this attribute is negated from + its opposite number by the frontend like Clang. The middle end or + backend should convert it back if possible. For example if -(a*b) + is converted to (-a)*b, the arg0 of `fmul` instruction should be + marked with `negated` attribute. ``alignstack(<n>)`` This indicates the alignment that should be considered by the backend when diff --git a/llvm/include/llvm/Bitcode/LLVMBitCodes.h b/llvm/include/llvm/Bitcode/LLVMBitCodes.h index 21fd27d9838db7..7e9d174db22026 100644 --- a/llvm/include/llvm/Bitcode/LLVMBitCodes.h +++ b/llvm/include/llvm/Bitcode/LLVMBitCodes.h @@ -788,6 +788,7 @@ enum AttributeKindCodes { ATTR_KIND_NO_EXT = 99, ATTR_KIND_NO_DIVERGENCE_SOURCE = 100, ATTR_KIND_SANITIZE_TYPE = 101, + ATTR_KIND_NEGATED = 102, }; enum ComdatSelectionKindCodes { diff --git a/llvm/include/llvm/IR/Attributes.td b/llvm/include/llvm/IR/Attributes.td index 61955cf883c3f1..baeca5d53f3c46 100644 --- a/llvm/include/llvm/IR/Attributes.td +++ b/llvm/include/llvm/IR/Attributes.td @@ -162,6 +162,9 @@ def Memory : IntAttr<"memory", IntersectCustom, [FnAttr]>; /// Forbidden floating-point classes. def NoFPClass : IntAttr<"nofpclass", IntersectCustom, [ParamAttr, RetAttr]>; +/// Converted from the opposite number +def Negated : EnumAttr<"negated", IntersectAnd, [ParamAttr, RetAttr]>; + /// Function must be optimized for size first. def MinSize : EnumAttr<"minsize", IntersectPreserve, [FnAttr]>; diff --git a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp index b4efd3928a2e6f..e87c9d2e13883d 100644 --- a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp +++ b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp @@ -755,6 +755,8 @@ static uint64_t getAttrKindEncoding(Attribute::AttrKind Kind) { return bitc::ATTR_KIND_MEMORY; case Attribute::NoFPClass: return bitc::ATTR_KIND_NOFPCLASS; + case Attribute::Negated: + return bitc::ATTR_KIND_NEGATED; case Attribute::Naked: return bitc::ATTR_KIND_NAKED; case Attribute::Nest: diff --git a/llvm/lib/Transforms/Utils/CodeExtractor.cpp b/llvm/lib/Transforms/Utils/CodeExtractor.cpp index 7ddb9e22c83441..4e1a8c560078aa 100644 --- a/llvm/lib/Transforms/Utils/CodeExtractor.cpp +++ b/llvm/lib/Transforms/Utils/CodeExtractor.cpp @@ -918,6 +918,7 @@ Function *CodeExtractor::constructFunctionDeclaration( case Attribute::PresplitCoroutine: case Attribute::Memory: case Attribute::NoFPClass: + case Attribute::Negated: case Attribute::CoroDestroyOnlyWhenComplete: case Attribute::CoroElideSafe: case Attribute::NoDivergenceSource: |
Contributor Author
| Depends on #121027 |
You can test this locally with the following command:git-clang-format --diff 4cb2a519db10f54815c8a4ccd5accbedc1cdfd07 9a8925b18e609ac646b2c16da81264a261545513 --extensions cpp,h,c -- clang/lib/CodeGen/CGExprScalar.cpp clang/test/CodeGen/constrained-math-builtins.c llvm/include/llvm/Bitcode/LLVMBitCodes.h llvm/lib/Bitcode/Writer/BitcodeWriter.cpp llvm/lib/Transforms/Utils/CodeExtractor.cppView the diff from clang-format here.diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp index 14d73de055..367ca42053 100644 --- a/clang/lib/CodeGen/CGExprScalar.cpp +++ b/clang/lib/CodeGen/CGExprScalar.cpp @@ -4121,9 +4121,11 @@ static Value* buildFMulAdd(llvm::Instruction *MulOp, Value *Addend, Addend->getType()), {MulOp0, MulOp1, Addend}); if (negMul) - dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(0, llvm::Attribute::Negated); + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(0, + llvm::Attribute::Negated); if (negAdd) - dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(2, llvm::Attribute::Negated); + dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(2, + llvm::Attribute::Negated); } else { FMulAdd = Builder.CreateCall( CGF.CGM.getIntrinsic(llvm::Intrinsic::fmuladd, Addend->getType()), |
arsenm requested changes Dec 26, 2024
Contributor
arsenm left a comment
arsenm left a comment There was a problem hiding this comment.
Can't really use an attribute for this. Attributes should not be used for adding constraints to an operation
| Addend->getType()), | ||
| {MulOp0, MulOp1, Addend}); | ||
| if (negMul) | ||
| dyn_cast<llvm::CallBase>(FMulAdd)->addParamAttr(0, llvm::Attribute::Negated); |
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
3 participants
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.
Use negated attribute if negMul or negAdd. So that we can lower
fneg+fmuladd to fmul+fsub if needed.
fneg/fmul/fadd vs fmul/fsub
c-a*bmay be different withc+(-a)*b.