[PATCH] D96547: [LoopVectorize] Let selectVectorizationFactor choose a scalable VF.

[Sander de Smalen via Phabricator via llvm-commits]

sdesmalen created this revision.
Herald added subscribers: bmahjour, hiraditya.
sdesmalen requested review of this revision.
Herald added a project: LLVM.

This patch changes the algorithm of `selectVectorizationFactor` to
iterate all power of 2 VFs between Min/MaxFactor, for both fixed
and scalable Min/MaxFactor VFs. It then invokes the cost-model as
normal, and selects the most cost-effective VF to vectorize the
loop.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D96547

Files:
  llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h
  llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
  llvm/test/Transforms/LoopVectorize/AArch64/scalable-vf-analysis.ll
  llvm/test/Transforms/LoopVectorize/AArch64/scalable-vf-hint.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D96547.323144.patch
Type: text/x-patch
Size: 8815 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210211/53c27cb8/attachment.bin>