This is an archive of the discontinued LLVM Phabricator instance.

[mlir][AMDGPU] Move to new buffer resource intrinsics

Authored by krzysz00 on Aug 3 2023, 4:04 PM.



The AMDGPU backend now has buffer resource intrinsics that take a ptr
addrspase (8) instead of a vector<4xi32>, improving LLVM's ability to
reason about their memory behavior. This commit moves MLIR to these
new functions.

Diff Detail

Event Timeline

krzysz00 created this revision.Aug 3 2023, 4:04 PM
Herald added a reviewer: dcaballe. · View Herald Transcript
Herald added a project: Restricted Project. · View Herald Transcript
krzysz00 requested review of this revision.Aug 3 2023, 4:04 PM
jsjodin accepted this revision.Sep 21 2023, 8:43 AM
jsjodin added a subscriber: jsjodin.

These changes look good to me.

This revision is now accepted and ready to land.Sep 21 2023, 8:43 AM
This revision was automatically updated to reflect the committed changes.