[PATCH] D67841: [SLP] avoid reduction transform on patterns that the backend can load-combine

[Alexey Bataev via Phabricator via llvm-commits]

ABataev added inline comments.


================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:6446
+  /// effectively impossible for the backend to undo.
+  bool getLoadCombineCost(TargetTransformInfo *TTI, Value *FirstReducedVal,
+                          unsigned ReduxWidth, int &ReduxCost) {
----------------
Maybe, better to do it a member of TargetTransformInfo rather than of SLP vectorizer?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D67841/new/

https://reviews.llvm.org/D67841