[Mlir-commits] [mlir] [mlir][amdgpu] Shared memory access optimization pass (PR #75627)

[Krzysztof Drewniak]

https://github.com/krzysz00 approved this pull request.

This looks reasonable and we'll clean this and its NVPTX equivalent up in a followup, it sounds like.

https://github.com/llvm/llvm-project/pull/75627