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

[Florian Hahn via llvm-commits]

================
@@ -99,6 +99,11 @@ struct VPlanTransforms {
   /// not valid.
   static bool adjustFixedOrderRecurrences(VPlan &Plan, VPBuilder &Builder);
 
+  /// Check if \p Plan contains any FMaxNumNoFMFs or FMinNumNoFMFs reductions.
+  /// If they do, try to update the vector loop to exit early if any input is
+  /// NaN and resume executing in the scalar loop to handle the NaNs there.
----------------
fhahn wrote:

updated, thanks

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