[PATCH] D92824: [SLP] Use the width of value truncated just before storing

[Anton Afanasyev via Phabricator via llvm-commits]

anton-afanasyev created this revision.
anton-afanasyev added reviewers: RKSimon, ABataev, spatel, xbolva00, dtemirbulatov.
Herald added a subscriber: hiraditya.
anton-afanasyev requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

For stores chain vectorization we choose the size of vector
elements to ensure we fit to minimum and maximum vector register
size for the number of elements given. This patch corrects vector
element size choosing the width of value truncated just before
storing instead of the width of value stored.

Fixes PR46983


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D92824

Files:
  llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
  llvm/test/Transforms/SLPVectorizer/X86/insert-after-bundle.ll
  llvm/test/Transforms/SLPVectorizer/X86/pr46983.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D92824.310115.patch
Type: text/x-patch
Size: 41739 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20201208/2ca8cc43/attachment.bin>