[PowerPC] Add missing pattern for VSX Scalar Negative Multiply-Subtract Single Precision

Summary:
This was found during review of https://reviews.llvm.org/D66050.
In the simple test of fdiv, we miss to fold
```
        fneg 2, 2
        xsmaddasp 3, 2, 0
```
to
```
        xsnmsubasp 3, 2, 0
```
We have the patterns for Double Precision and vectors, just missing
Single Precision, the patch add that.

Reviewers: #powerpc, hfinkel, nemanjai, steven.zhang

Reviewed By: #powerpc, steven.zhang

Subscribers: wuzish, hiraditya, kbarton, MaskRay, shchenz, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D67595

llvm-svn: 372985

diff --git a/llvm/lib/Target/PowerPC/PPCInstrVSX.td b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
index 910d98d..bf15070 100644 (file)
--- a/llvm/lib/Target/PowerPC/PPCInstrVSX.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
@@ -1512,6 +1512,12 @@ let AddedComplexity = 400 in { // Prefer VSX patterns over non-VSX patterns.
                           AltVSXFMARel;
   }
 
+  // Additional xsnmsubasp patterns: -a*b + c == -(a*b - c)
+  def : Pat<(fma (fneg f32:$A), f32:$B, f32:$C),
+            (XSNMSUBASP $C, $A, $B)>;
+  def : Pat<(fma f32:$A, (fneg f32:$C), f32:$B),
+            (XSNMSUBASP $C, $A, $B)>;
+
   // Single Precision Conversions (FP <-> INT)
   def XSCVSXDSP : XX2Form<60, 312,
                       (outs vssrc:$XT), (ins vsfrc:$XB),

diff --git a/llvm/test/CodeGen/PowerPC/fdiv.ll b/llvm/test/CodeGen/PowerPC/fdiv.ll
index 4e73dd2..183e514 100644 (file)
--- a/llvm/test/CodeGen/PowerPC/fdiv.ll
+++ b/llvm/test/CodeGen/PowerPC/fdiv.ll
@@ -5,9 +5,8 @@ define dso_local float @foo(float %0, float %1) local_unnamed_addr {
 ; CHECK-LABEL: foo:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    xsresp 3, 2
-; CHECK-NEXT:    fneg 2, 2
 ; CHECK-NEXT:    xsmulsp 0, 1, 3
-; CHECK-NEXT:    xsmaddasp 1, 2, 0
+; CHECK-NEXT:    xsnmsubasp 1, 2, 0
 ; CHECK-NEXT:    xsmaddasp 0, 3, 1
 ; CHECK-NEXT:    fmr 1, 0
 ; CHECK-NEXT:    blr