[PATCH] D62871: [Codegen] (X & (C l>>/<< Y)) ==/!= 0 --> ((X <</l>> Y) & C) ==/!= 0 fold

[Nicolai Hähnle via Phabricator via llvm-commits]

nhaehnle added a comment.

FWIW, the AMDGPU changes aren't //entirely// neutral because they shift the wait for previous memory loads up by one instructions, so it's a minor loss. However, that seems benign enough that I wouldn't object.


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D62871/new/

https://reviews.llvm.org/D62871