[PATCH] D132740: [SLP]Improve lookup of the buildvector top insertelement instruction.

[Alexey Bataev via Phabricator via llvm-commits]

ABataev created this revision.
ABataev added reviewers: RKSimon, vdmitrie.
Herald added subscribers: vporpo, hiraditya.
Herald added a project: All.
ABataev requested review of this revision.
Herald added a subscriber: pcwang-thead.
Herald added a project: LLVM.

When estimating the cost of the in-tree vectorized scalars in
buildvector sequences, need to take into account the vectorized
insertelement instruction. The top of the buildvector seuences is the
topmost vectorized insertelement instruction, because it will have

> than 1 use after the vectorization.

For the affected test case improves througput from 21 to 16 (per
llvm-mca).


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D132740

Files:
  llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
  llvm/test/Transforms/SLPVectorizer/X86/vec_list_bias_external_insert_shuffled.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D132740.455894.patch
Type: text/x-patch
Size: 7343 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220826/1f289a64/attachment.bin>