[PATCH] D110995: [X86] combineMulToPMADDWD - handle any pow2 vector type and split to legal types

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon updated this revision to Diff 384720.
RKSimon added a comment.

Fix pre-SSE41 regressions - if both operands are sext from i8 (or smaller) pmull/pmulh are still better - if they are a sext/zext mix we are better off using pmaddwd.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110995/new/

https://reviews.llvm.org/D110995

Files:
  llvm/lib/Target/X86/X86ISelLowering.cpp
  llvm/test/CodeGen/X86/pmaddubsw.ll
  llvm/test/CodeGen/X86/pmulh.ll
  llvm/test/CodeGen/X86/shrink_vmul.ll
  llvm/test/CodeGen/X86/slow-pmulld.ll
  llvm/test/CodeGen/X86/vshli-simplify-demanded-bits.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D110995.384720.patch
Type: text/x-patch
Size: 44387 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20211104/438338f9/attachment.bin>