[PATCH] D100121: [LV] Let selectVectorizationFactor reason directly on VectorizationFactor.

[Sander de Smalen via Phabricator via llvm-commits]

sdesmalen added inline comments.


================
Comment at: llvm/lib/Transforms/Vectorize/LoopVectorize.cpp:1607
+  /// that of B.
+  bool isMoreProfitable(const VectorizationFactor &A,
+                        const VectorizationFactor &B) const;
----------------
bmahjour wrote:
> I think this should become a member of the `VectorizationFactor` struct....then code like `A.isMoreProfitableThan(B)` reads better.
I'm happy to implement your suggestion, although I think in the future a `LoopVectorizationCostModel::isMoreProfitable` method will be needed anyway, so that the method can query the LoopHints to understand how to interpret the costs of a scalable VF, and possibly favour it over fixed-width VFs of similar cost. But perhaps that's just something to worry about in a future patch?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D100121/new/

https://reviews.llvm.org/D100121