[llvm] [LV] Vectorize maxnum/minnum w/o fast-math flags. (PR #148239)

[Florian Hahn via llvm-commits]

================
@@ -84,6 +84,7 @@ Type *VPTypeAnalysis::inferScalarTypeForRecipe(const VPInstruction *R) {
     return ResTy;
   }
   case Instruction::ICmp:
+  case Instruction::FCmp:
----------------
fhahn wrote:

Yep, only this patch introduces FCmp VPInstructions, so only tested now.

https://github.com/llvm/llvm-project/pull/148239