[PATCH] D158750: New intrinsic void llvm.amdgcn.s.nop(i16)
David Stuttard via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Sep 5 01:49:03 PDT 2023
dstuttard added a comment.
In D158750#4637244 <https://reviews.llvm.org/D158750#4637244>, @foad wrote:
> Depending on the use case, you might also have to mark it as a scheduling barrier, not just having side effects.
It is possible that might be required, but the behaviour is modelled on s_sleep which only tags "hasSideEffects" and doesn't attempt to make it a scheduling barrier as well.
I propose to leave the implementation as-is and it can be re-visited if a scheduling barrier needs to be enforced.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D158750/new/
https://reviews.llvm.org/D158750
More information about the llvm-commits
mailing list