[Openmp-commits] [PATCH] D92463: [libomptarget][AMDGPU] Remove MaxParallelLevel

[Jon Chesterfield via Phabricator via Openmp-commits]

JonChesterfield accepted this revision.
JonChesterfield added a comment.
This revision is now accepted and ready to land.

Very good, thank you.

Some context. Nested parallel is implemented by a transform to sequential execution. This transform uses linked call frames, allocated by malloc from the gpu within deviceRTL.

A previous version of amdgpu's deviceRTL had a buggy malloc implementation. The code that this patch deletes worked around that by allocating the gpu memory on the host instead.

That work around is now gone from the downstream fork (because malloc is no longer buggy), thus also removed from here.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D92463/new/

https://reviews.llvm.org/D92463