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

[Florian Hahn via llvm-commits]

================
@@ -47,6 +47,8 @@ enum class RecurKind {
   FMul,     ///< Product of floats.
   FMin,     ///< FP min implemented in terms of select(cmp()).
   FMax,     ///< FP max implemented in terms of select(cmp()).
+  FMinNum, ///< FP min with llvm.minnum semantics including NaNs.
+  FMaxNum, ///< FP max with llvm.maxnum semantics including NaNs.
----------------
fhahn wrote:

Done thanks

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