[PATCH] D115283: [AMDGPU] Set "amdgpu_hostcall" module flag if an AMDGPU function has calls to device lib functions that use hostcalls.

Jon Chesterfield via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Fri Jan 21 10:22:14 PST 2022


JonChesterfield added a comment.

The asm variable used by rocm aomp is zero overhead, needs no compiler support and works exactly as one would wish under inlining or code elimination. The main argument against that approach seems to be it's an abi break, much like this patch was, and that it is per-code-object instead of per-function, which I still think is a benefit.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D115283/new/

https://reviews.llvm.org/D115283



More information about the cfe-commits mailing list