[PATCH] D158724: [AArch64][LoopVectorize] Add truncated store values to list of types for widening

[Irina Dobrescu via Phabricator via llvm-commits]

Rin created this revision.
Herald added subscribers: artagnon, shiva0217, sunshaoce, hiraditya, kristof.beyls.
Herald added a project: All.
Rin requested review of this revision.
Herald added subscribers: llvm-commits, wangpc.
Herald added a project: LLVM.

When dealing with truncated stores, the Min/Max width would not consider source values of truncated stores. Instead it would only look at the type value after performing the truncate instruction. This patch adds the type value of truncated stores to the list of types for widening.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D158724

Files:
  llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
  llvm/test/Transforms/LoopVectorize/AArch64/deterministic-type-shrinkage.ll
  llvm/test/Transforms/LoopVectorize/AArch64/extend-vectorization-factor-for-unprofitable-memops.ll
  llvm/test/Transforms/LoopVectorize/AArch64/loop-vectorization-factors.ll
  llvm/test/Transforms/LoopVectorize/AArch64/outer_loop_test1_no_explicit_vect_width.ll
  llvm/test/Transforms/LoopVectorize/AArch64/pr60831-sve-inv-store-crash.ll
  llvm/test/Transforms/LoopVectorize/AArch64/truncate-type-widening.ll
  llvm/test/Transforms/LoopVectorize/AArch64/type-shrinkage-zext-costs.ll
  llvm/test/Transforms/LoopVectorize/optimal-epilog-vectorization.ll
  llvm/test/Transforms/LoopVectorize/vplan-stress-test-no-explict-vf.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D158724.553073.patch
Type: text/x-patch
Size: 74097 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230824/c58b51a5/attachment-0001.bin>