[PATCH] D153744: [LoopUnroll] adjust for new `convergent` semantics

[Jon Chesterfield via Phabricator via llvm-commits]

JonChesterfield added a comment.

This looks quite high risk. How is it tested beyond this lit test, specifically with code that runs on the GPU? It's really easy to get broken transforms to pass a lit test where the checks were generated from the implementation.

Also could you point reviewer's to the documentation on the heart intrinsic system and how it relates to the current modeling? I know there's a slide deque or two somewhere, hoping for something more formal.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D153744/new/

https://reviews.llvm.org/D153744