[llvm] [X86] X86FixupInstTuning - prefer VPBLENDD to VPBLENDW shuffles on AVX2+ targets (PR #144269)

[Simon Pilgrim via llvm-commits]

================
@@ -270,6 +290,15 @@ bool X86FixupInstTuningPass::processInstruction(
     return ProcessBLENDToMOV(X86::VMOVSSrr, 0xF, 0x1) ||
            ProcessBLENDToMOV(X86::VMOVSDrr, 0xF, 0x3);
 
+  case X86::VPBLENDWrri:
+    return ProcessBLENDWToBLENDD(X86::VPBLENDDrri, 4);
+  case X86::VPBLENDWrmi:
+    return ProcessBLENDWToBLENDD(X86::VPBLENDDrmi, 8);
----------------
RKSimon wrote:

Nice catch! I'm going to drop them until we have test coverage

https://github.com/llvm/llvm-project/pull/144269