[PATCH] D110372: [AMDGPU] Limit promote alloca max size in functions

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added a comment.
Herald added a subscriber: kosarev.
Herald added a project: All.

> There is no reason to eliminate scratch access to get

another scratch access instead.

This can be beneficial since one pair CSR spills happen once if this helps avoid stack access inside a loop. It requires considering additional context to know this though


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110372/new/

https://reviews.llvm.org/D110372