[X86][SSE] Fold add(shuffle(),shuffle()) to hadd on 'slow' targets (PR39920)

As reported on PR39920, "slow horizontal ops" targets tend to internally expand to 2*shuffle+add/sub - so if we can reduce 2*shuffle+add/sub to a hadd/sub then we should do it - similar port usage but reduced instruction count.

This works out in most cases, although the "PR22377" regression in vector-shuffle-combining.ll is annoying - going from 2*shuffle+add+shuffle to hadd+2*shuffle - I've opened PR41813 to cover this.

Differential Revision: https://reviews.llvm.org/D61308

llvm-svn: 360360

1 files changed, 3 insertions, 2 deletions

diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index 26619571821..fa3f61dc8c6 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -39412,7 +39412,8 @@ static bool isHorizontalBinOp(SDValue &LHS, SDValue &RHS, SelectionDAG &DAG,
   GetShuffle(RHS, C, D, RMask);
 
   // At least one of the operands should be a vector shuffle.
-  if (LMask.empty() && RMask.empty())
+  unsigned NumShuffles = (LMask.empty() ? 0 : 1) + (RMask.empty() ? 0 : 1);
+  if (NumShuffles == 0)
     return false;
 
   if (LMask.empty()) {
@@ -39474,7 +39475,7 @@ static bool isHorizontalBinOp(SDValue &LHS, SDValue &RHS, SelectionDAG &DAG,
   LHS = A.getNode() ? A : B; // If A is 'UNDEF', use B for it.
   RHS = B.getNode() ? B : A; // If B is 'UNDEF', use A for it.
 
-  if (!shouldUseHorizontalOp(LHS == RHS, DAG, Subtarget))
+  if (!shouldUseHorizontalOp(LHS == RHS && NumShuffles < 2, DAG, Subtarget))
     return false;
 
   LHS = DAG.getBitcast(VT, LHS);