[PATCH] D14909: [X86][FMA] Optimize FNEG(FMUL) Patterns

[Simon Pilgrim via llvm-commits]

RKSimon updated this revision to Diff 41140.
RKSimon added a comment.

Updated with Sanjay & Elena's comments


Repository:
  rL LLVM

http://reviews.llvm.org/D14909

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/CodeGen/X86/fma_patterns.ll

Index: test/CodeGen/X86/fma_patterns.ll
===================================================================
--- test/CodeGen/X86/fma_patterns.ll
+++ test/CodeGen/X86/fma_patterns.ll
@@ -667,4 +667,40 @@
   ret <4 x float> %a
 }
 
+; (fneg (fmul x, y)) -> (fnmadd x, y, 0)
+
+define double @test_f64_fneg_fmul(double %x, double %y) #0 {
+; CHECK_FMA-LABEL: test_f64_fneg_fmul:
+; CHECK_FMA:       # BB#0:
+; CHECK_FMA-NEXT:    vxorps %xmm2, %xmm2, %xmm2
+; CHECK_FMA-NEXT:    vfnmadd213sd %xmm2, %xmm1, %xmm0
+; CHECK_FMA-NEXT:    retq
+;
+; CHECK_FMA4-LABEL: test_f64_fneg_fmul:
+; CHECK_FMA4:       # BB#0:
+; CHECK_FMA4-NEXT:    vxorps %xmm2, %xmm2, %xmm2
+; CHECK_FMA4-NEXT:    vfnmaddsd %xmm2, %xmm1, %xmm0, %xmm0
+; CHECK_FMA4-NEXT:    retq
+  %m = fmul double %x, %y
+  %n = fsub double -0.0, %m
+  ret double %n
+}
+
+define <4 x float> @test_v4f32_fneg_fmul(<4 x float> %x, <4 x float> %y) #0 {
+; CHECK_FMA-LABEL: test_v4f32_fneg_fmul:
+; CHECK_FMA:       # BB#0:
+; CHECK_FMA-NEXT:    vxorps %xmm2, %xmm2, %xmm2
+; CHECK_FMA-NEXT:    vfnmadd213ps %xmm2, %xmm1, %xmm0
+; CHECK_FMA-NEXT:    retq
+;
+; CHECK_FMA4-LABEL: test_v4f32_fneg_fmul:
+; CHECK_FMA4:       # BB#0:
+; CHECK_FMA4-NEXT:    vxorps %xmm2, %xmm2, %xmm2
+; CHECK_FMA4-NEXT:    vfnmaddps %xmm2, %xmm1, %xmm0, %xmm0
+; CHECK_FMA4-NEXT:    retq
+  %m = fmul <4 x float> %x, %y
+  %n = fsub <4 x float> <float -0.0, float -0.0, float -0.0, float -0.0>, %m
+  ret <4 x float> %n
+}
+
 attributes #0 = { "unsafe-fp-math"="true" }
Index: lib/Target/X86/X86ISelLowering.cpp
===================================================================
--- lib/Target/X86/X86ISelLowering.cpp
+++ lib/Target/X86/X86ISelLowering.cpp
@@ -26174,24 +26174,35 @@
 static SDValue PerformFNEGCombine(SDNode *N, SelectionDAG &DAG,
                                   const X86Subtarget *Subtarget) {
   EVT VT = N->getValueType(0);
+  EVT SVT = VT.getScalarType();
   SDValue Arg = N->getOperand(0);
+  SDLoc DL(N);
+
+  // If we're negating a FMUL node on a target with FMA, then we can avoid the
+  // use of a constant by performing (0 - A*B) instead.
+  if (Arg.getOpcode() == ISD::FMUL && (SVT == MVT::f32 || SVT == MVT::f64) &&
+      (Subtarget->hasFMA() || Subtarget->hasFMA4())) {
+    SDValue Zero = DAG.getConstantFP(0.0, DL, VT);
+    return DAG.getNode(X86ISD::FNMADD, DL, VT, Arg.getOperand(0),
+                       Arg.getOperand(1), Zero);
+  }
 
   // If we're negating a FMA node, then we can adjust the
   // instruction to include the extra negation.
   if (Arg.hasOneUse()) {
     switch (Arg.getOpcode()) {
-      case X86ISD::FMADD:
-        return DAG.getNode(X86ISD::FNMSUB, SDLoc(N), VT, Arg.getOperand(0),
-                           Arg.getOperand(1), Arg.getOperand(2));
-      case X86ISD::FMSUB:
-        return DAG.getNode(X86ISD::FNMADD, SDLoc(N), VT, Arg.getOperand(0),
-                           Arg.getOperand(1), Arg.getOperand(2));
-      case X86ISD::FNMADD:
-        return DAG.getNode(X86ISD::FMSUB, SDLoc(N), VT, Arg.getOperand(0),
-                           Arg.getOperand(1), Arg.getOperand(2));
-      case X86ISD::FNMSUB:
-        return DAG.getNode(X86ISD::FMADD, SDLoc(N), VT, Arg.getOperand(0),
-                           Arg.getOperand(1), Arg.getOperand(2));
+    case X86ISD::FMADD:
+      return DAG.getNode(X86ISD::FNMSUB, DL, VT, Arg.getOperand(0),
+                         Arg.getOperand(1), Arg.getOperand(2));
+    case X86ISD::FMSUB:
+      return DAG.getNode(X86ISD::FNMADD, DL, VT, Arg.getOperand(0),
+                         Arg.getOperand(1), Arg.getOperand(2));
+    case X86ISD::FNMADD:
+      return DAG.getNode(X86ISD::FMSUB, DL, VT, Arg.getOperand(0),
+                         Arg.getOperand(1), Arg.getOperand(2));
+    case X86ISD::FNMSUB:
+      return DAG.getNode(X86ISD::FMADD, DL, VT, Arg.getOperand(0),
+                         Arg.getOperand(1), Arg.getOperand(2));
     }
   }
   return SDValue();


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D14909.41140.patch
Type: text/x-patch
Size: 4018 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151125/276c4d02/attachment.bin>