[PATCH] D155267: [DAGCombiner] Change foldAndOrOfSETCC() to optimize and/or patterns with floating points.

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp:6055
+  bool isNotNaN =
+      DAG.isKnownNeverNaN(Operand1) && DAG.isKnownNeverNaN(Operand2);
+  bool isNotSNaN =
----------------
Check operand 2 first, it's canonically simpler


================
Comment at: llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp:6087-6092
+      if (isNotSNaN)
+        return ISD::FMINNUM_IEEE;
+
+      if (isFMINNUM_FMAXNUM_Supported)
+        return ISD::FMINNUM;
+
----------------
arsenm wrote:
> Can this be reordered to try FMINNUM/FMAXNUM first?
The point to f this was to try to avoid recursing 6 times to call isKnownNeverSNan, you can try to avoid that in the case where you don't need to worry about it


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D155267/new/

https://reviews.llvm.org/D155267