[PATCH] D63226: AMDGPU: Fold readlane intrinsics of constants

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added a comment.

In D63226#1540466 <https://reviews.llvm.org/D63226#1540466>, @rampitec wrote:

> How this can even happen? Someone needs to use an intrinsic on a literal? This use doesn't make a lot of sense. Do you have a real example of such use?


You can use readfirstlane to effectively assert that a function argument must be uniform. After inlining, it might see a constant


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D63226/new/

https://reviews.llvm.org/D63226