[llvm] [X86] Attempt to use VPMADD52L/VPMULUDQ instead of VPMULLQ on slow VPMULLQ targets (or when VPMULLQ is unavailable) (PR #171760)

[via llvm-commits]

================
@@ -49926,6 +49873,41 @@ static SDValue combineMul(SDNode *N, SelectionDAG &DAG,
   if (SDValue V = combineMulToPMULDQ(N, DL, DAG, Subtarget))
     return V;
 
+  if (VT.getScalarType() == MVT::i64 && Subtarget.isPMULLQSlow()) {
----------------
houngkoungting wrote:

I'm back. You were right about the helper function.I've updated the commit . Please check it when you have time.

https://github.com/llvm/llvm-project/pull/171760