[PATCH] D60318: [ExpandMemCmp] Improve generated code for simple non-equality compares.

[Sanjay Patel via Phabricator via llvm-commits]

spatel added a comment.

Alternatively, what do you think about making ExpandMemCmp a late IR optimization pass like the vectorizer passes?
When I was looking at memcmp optimizations initially, it was clear that in some larger examples we would benefit from running instcombine/CSE after the expansion. By moving this pass up, we wouldn't have to worry about adding optimizations (like this patch) to the backend because it will all be done for us in existing IR optimization passes.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D60318/new/

https://reviews.llvm.org/D60318