[PATCH] D69044: [X86] Allow up to 4 loads per inline memcmp()

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

In D69044#1715641 <https://reviews.llvm.org/D69044#1715641>, @davezarzycki wrote:

> Oh, and the memcmp() expansion only honors -mprefer-vector-width if the width is 256 or 512. I don't think any part of the X86 code gen actually honors 128 when 256 or 512 is possible.


Nothing disables 256-bit types the way we do for 512. But X86TargetLowering::getOptimalMemOpType() honors -mprefer-vector-width=128


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D69044/new/

https://reviews.llvm.org/D69044