[Mlir-commits] [mlir] [mlir][gpu] Pattern to promote `gpu.shuffle` to specialized AMDGPU ops (PR #137109)

[Krzysztof Drewniak]

krzysz00 wrote:

Independent of `gpu.subgroup_reduce` there might be `gpu.shuffle` cases that can be lowered more efficiently, perhaps

https://github.com/llvm/llvm-project/pull/137109