[InstCombine] fold fneg into constant operand of fmul/fdiv

[spatel]

Summary

I think this is the missing IR fold noted in D50195, but AFAICT, we don't need any fast-math to enable the transform. FP negation can always be folded into an fmul/fdiv constant to eliminate the fneg.

I've limited this to one-use to ensure that we are eliminating an instruction rather than replacing fneg by a potentially expensive fdiv or fmul.

[mcberg2017]

Looked it over in detail, the transformations all track.
LGTM, up to you if you want second set of eyes.

[spatel]

Marking as accepted here on Phab to avoid unknown state when committing.

llvm/trunk/lib/Transforms/InstCombine/InstCombineAddSub.cpp

if (Instruction *X = foldShuffledBinop(I))
     return X;
 
   // Subtraction from -0.0 is the canonical form of fneg.
   // fsub nsz 0, X ==> fsub nsz -0.0, X
   Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);
   if (I.hasNoSignedZeros() && match(Op0, m_PosZeroFP()))
     return BinaryOperator::CreateFNegFMF(Op1, &I);
 
+  Value *X, *Y;
+  Constant *C;
+
+  // Fold negation into constant operand. This is limited with one-use because
+  // fneg is assumed better for analysis and cheaper in codegen than fmul/fdiv.
+  // -(X * C) --> X * (-C)
+  if (match(&I, m_FNeg(m_OneUse(m_FMul(m_Value(X), m_Constant(C))))))
+    return BinaryOperator::CreateFMulFMF(X, ConstantExpr::getFNeg(C), &I);
+  // -(X / C) --> X / (-C)
+  if (match(&I, m_FNeg(m_OneUse(m_FDiv(m_Value(X), m_Constant(C))))))
+    return BinaryOperator::CreateFDivFMF(X, ConstantExpr::getFNeg(C), &I);
+  // -(C / X) --> (-C) / X
+  if (match(&I, m_FNeg(m_OneUse(m_FDiv(m_Constant(C), m_Value(X))))))
+    return BinaryOperator::CreateFDivFMF(ConstantExpr::getFNeg(C), X, &I);
+
   // If Op0 is not -0.0 or we can ignore -0.0: Z - (X - Y) --> Z + (Y - X)
   // Canonicalize to fadd to make analysis easier.
   // This can also help codegen because fadd is commutative.
   // Note that if this fsub was really an fneg, the fadd with -0.0 will get
   // killed later. We still limit that particular transform with 'hasOneUse'
   // because an fneg is assumed better/cheaper than a generic fsub.
-  Value *X, *Y;
   if (I.hasNoSignedZeros() || CannotBeNegativeZero(Op0, SQ.TLI)) {
     if (match(Op1, m_OneUse(m_FSub(m_Value(X), m_Value(Y))))) {
       Value *NewSub = Builder.CreateFSubFMF(Y, X, &I);
       return BinaryOperator::CreateFAddFMF(Op0, NewSub, &I);
     }
   }
 
   if (isa<Constant>(Op0))
     if (SelectInst *SI = dyn_cast<SelectInst>(Op1))
       if (Instruction *NV = FoldOpIntoSelect(I, SI))
         return NV;
 
   // X - C --> X + (-C)
   // But don't transform constant expressions because there's an inverse fold
   // for X + (-Y) --> X - Y.
-  Constant *C;
   if (match(Op1, m_Constant(C)) && !isa<ConstantExpr>(Op1))
     return BinaryOperator::CreateFAddFMF(Op0, ConstantExpr::getFNeg(C), &I);
 
   // X - (-Y) --> X + Y
   if (match(Op1, m_FNeg(m_Value(Y))))
     return BinaryOperator::CreateFAddFMF(Op0, Y, &I);
 
   // Similar to above, but look through a cast of the negated value:

llvm/trunk/test/Transforms/InstCombine/fneg.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: opt < %s -instcombine -S | FileCheck %s
 
 declare void @use(float)
 
 ; -(X * C) --> X * (-C)
 
 define float @fmul_fneg(float %x) {
 ; CHECK-LABEL: @fmul_fneg(
-; CHECK-NEXT:    [[M:%.*]] = fmul float [[X:%.*]], 4.200000e+01
-; CHECK-NEXT:    [[R:%.*]] = fsub float -0.000000e+00, [[M]]
+; CHECK-NEXT:    [[R:%.*]] = fmul float [[X:%.*]], -4.200000e+01
 ; CHECK-NEXT:    ret float [[R]]
 ;
   %m = fmul float %x, 42.0
   %r = fsub float -0.0, %m
   ret float %r
 }
 
 ; Fast math is not required, but it should be propagated.
 
 define float @fmul_fneg_fmf(float %x) {
 ; CHECK-LABEL: @fmul_fneg_fmf(
-; CHECK-NEXT:    [[M:%.*]] = fmul float [[X:%.*]], 4.200000e+01
-; CHECK-NEXT:    [[R:%.*]] = fsub reassoc nsz float -0.000000e+00, [[M]]
+; CHECK-NEXT:    [[R:%.*]] = fmul reassoc nsz float [[X:%.*]], -4.200000e+01
 ; CHECK-NEXT:    ret float [[R]]
 ;
   %m = fmul float %x, 42.0
   %r = fsub reassoc nsz float -0.0, %m
   ret float %r
 }
 
 ; Extra use prevents the fold. We don't want to replace the fneg with an fmul.
;
   call void @use(float %m)
   ret float %r
 }
 
 ; Try a vector. Use special constants (NaN, INF, undef) because they don't change anything.
 
 define <4 x double> @fmul_fneg_vec(<4 x double> %x) {
 ; CHECK-LABEL: @fmul_fneg_vec(
-; CHECK-NEXT:    [[M:%.*]] = fmul <4 x double> [[X:%.*]], <double 4.200000e+01, double 0xFF80000000000000, double 0x7FF0000000000000, double undef>
-; CHECK-NEXT:    [[R:%.*]] = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, [[M]]
+; CHECK-NEXT:    [[R:%.*]] = fmul <4 x double> [[X:%.*]], <double -4.200000e+01, double 0x7F80000000000000, double 0xFFF0000000000000, double 0x7FF8000000000000>
 ; CHECK-NEXT:    ret <4 x double> [[R]]
 ;
   %m = fmul <4 x double> %x, <double 42.0, double 0x7FF80000000000000, double 0x7FF0000000000000, double undef>
   %r = fsub <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %m
   ret <4 x double> %r
 }
 
 ; -(X / C) --> X / (-C)
 
 define float @fdiv_op1_constant_fneg(float %x) {
 ; CHECK-LABEL: @fdiv_op1_constant_fneg(
-; CHECK-NEXT:    [[D:%.*]] = fdiv float [[X:%.*]], -4.200000e+01
-; CHECK-NEXT:    [[R:%.*]] = fsub float -0.000000e+00, [[D]]
+; CHECK-NEXT:    [[R:%.*]] = fdiv float [[X:%.*]], 4.200000e+01
 ; CHECK-NEXT:    ret float [[R]]
 ;
   %d = fdiv float %x, -42.0
   %r = fsub float -0.0, %d
   ret float %r
 }
 
 ; Fast math is not required, but it should be propagated.
 
 define float @fdiv_op1_constant_fneg_fmf(float %x) {
 ; CHECK-LABEL: @fdiv_op1_constant_fneg_fmf(
-; CHECK-NEXT:    [[D:%.*]] = fdiv float [[X:%.*]], -4.200000e+01
-; CHECK-NEXT:    [[R:%.*]] = fsub nnan float -0.000000e+00, [[D]]
+; CHECK-NEXT:    [[R:%.*]] = fdiv nnan float [[X:%.*]], 4.200000e+01
 ; CHECK-NEXT:    ret float [[R]]
 ;
   %d = fdiv float %x, -42.0
   %r = fsub nnan float -0.0, %d
   ret float %r
 }
 
 ; Extra use prevents the fold. We don't want to replace the fneg with an fdiv.
;
   call void @use(float %d)
   ret float %r
 }
 
 ; Try a vector. Use special constants (NaN, INF, undef) because they don't change anything.
 
 define <4 x double> @fdiv_op1_constant_fneg_vec(<4 x double> %x) {
 ; CHECK-LABEL: @fdiv_op1_constant_fneg_vec(
-; CHECK-NEXT:    [[D:%.*]] = fdiv <4 x double> [[X:%.*]], <double -4.200000e+01, double 0xFFF800000ABCD000, double 0xFFF0000000000000, double undef>
-; CHECK-NEXT:    [[R:%.*]] = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, [[D]]
+; CHECK-NEXT:    [[R:%.*]] = fdiv <4 x double> [[X:%.*]], <double 4.200000e+01, double 0x7FF800000ABCD000, double 0x7FF0000000000000, double 0x7FF8000000000000>
 ; CHECK-NEXT:    ret <4 x double> [[R]]
 ;
   %d = fdiv <4 x double> %x, <double -42.0, double 0xFFF800000ABCD000, double 0xFFF0000000000000, double undef>
   %r = fsub <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %d
   ret <4 x double> %r
 }
 
 ; -(C / X) --> (-C) / X
 
 define float @fdiv_op0_constant_fneg(float %x) {
 ; CHECK-LABEL: @fdiv_op0_constant_fneg(
-; CHECK-NEXT:    [[D:%.*]] = fdiv float 4.200000e+01, [[X:%.*]]
-; CHECK-NEXT:    [[R:%.*]] = fsub float -0.000000e+00, [[D]]
+; CHECK-NEXT:    [[R:%.*]] = fdiv float -4.200000e+01, [[X:%.*]]
 ; CHECK-NEXT:    ret float [[R]]
 ;
   %d = fdiv float 42.0, %x
   %r = fsub float -0.0, %d
   ret float %r
 }
 
 ; Fast math is not required, but it should be propagated.
 
 define float @fdiv_op0_constant_fneg_fmf(float %x) {
 ; CHECK-LABEL: @fdiv_op0_constant_fneg_fmf(
-; CHECK-NEXT:    [[D:%.*]] = fdiv float 4.200000e+01, [[X:%.*]]
-; CHECK-NEXT:    [[R:%.*]] = fsub fast float -0.000000e+00, [[D]]
+; CHECK-NEXT:    [[R:%.*]] = fdiv fast float -4.200000e+01, [[X:%.*]]
 ; CHECK-NEXT:    ret float [[R]]
 ;
   %d = fdiv float 42.0, %x
   %r = fsub fast float -0.0, %d
   ret float %r
 }
 
 ; Extra use prevents the fold. We don't want to replace the fneg with an fdiv.
;
   call void @use(float %d)
   ret float %r
 }
 
 ; Try a vector. Use special constants (NaN, INF, undef) because they don't change anything.
 
 define <4 x double> @fdiv_op0_constant_fneg_vec(<4 x double> %x) {
 ; CHECK-LABEL: @fdiv_op0_constant_fneg_vec(
-; CHECK-NEXT:    [[D:%.*]] = fdiv <4 x double> <double -4.200000e+01, double 0xFF80000000000000, double 0xFFF0000000000000, double undef>, [[X:%.*]]
-; CHECK-NEXT:    [[R:%.*]] = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, [[D]]
+; CHECK-NEXT:    [[R:%.*]] = fdiv <4 x double> <double 4.200000e+01, double 0x7F80000000000000, double 0x7FF0000000000000, double 0x7FF8000000000000>, [[X:%.*]]
 ; CHECK-NEXT:    ret <4 x double> [[R]]
 ;
   %d = fdiv <4 x double> <double -42.0, double 0x7FF80000000000000, double 0xFFF0000000000000, double undef>, %x
   %r = fsub <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %d
   ret <4 x double> %r
 }
 

llvm/trunk/test/Transforms/InstCombine/fsub.ll

 ; CHECK-NEXT:    [[T2:%.*]] = fadd nsz float [[TMP1]], [[Z:%.*]]
 ; CHECK-NEXT:    ret float [[T2]]
 ;
   %t1 = fsub float %x, %y
   %t2 = fsub nsz float %z, %t1
   ret float %t2
 }
 
-; With nsz and reassoc: Y - (X + Y) --> -X 
+; With nsz and reassoc: Y - ((X * 5) + Y) --> X * -5
+
 define float @sub_add_neg_x(float %x, float %y) {
 ; CHECK-LABEL: @sub_add_neg_x(
-; CHECK-NEXT:    [[TMP1:%.*]] = fmul float [[X:%.*]], 5.000000e+00
-; CHECK-NEXT:    [[T2:%.*]] = fsub reassoc nsz float -0.000000e+00, [[TMP1]]
-; CHECK-NEXT:    ret float [[T2]]
+; CHECK-NEXT:    [[TMP1:%.*]] = fmul reassoc nsz float [[X:%.*]], -5.000000e+00
+; CHECK-NEXT:    ret float [[TMP1]]
 ;
   %mul = fmul float %x, 5.000000e+00
-  %add = fadd float %mul, %y 
+  %add = fadd float %mul, %y
   %r = fsub nsz reassoc float %y, %add
   ret float %r
 }
 
 ; Same as above: if 'Z' is not -0.0, swap fsub operands and convert to fadd.
 
 define float @sub_sub_known_not_negzero(float %x, float %y) {
 ; CHECK-LABEL: @sub_sub_known_not_negzero(