[PATCH] D107286: [LV] Widen ExtractValue instructions instead of scalarizing.

Sander de Smalen via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Aug 2 09:22:57 PDT 2021


sdesmalen created this revision.
Herald added a subscriber: hiraditya.
sdesmalen requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

There is no particular reason that ExtractValue must be scalarized,
since it's operands must always be scalar or loop-invariant.

This is additionally a requirement for scalable vectors, which cannot
scalarize such operations at the moment.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D107286

Files:
  llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
  llvm/test/Transforms/LoopVectorize/AArch64/extractvalue-no-scalarization-required.ll
  llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-extractvalue.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D107286.363492.patch
Type: text/x-patch
Size: 7838 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210802/6703d4cd/attachment.bin>


More information about the llvm-commits mailing list