[Mlir-commits] [mlir] [MLIR][ROCDL] Remove patterns for ops supported as intrinsics in the AMDGPU backend (PR #102971)

[Krzysztof Drewniak]

krzysz00 wrote:

Ok, so maybe I'm misreading the above comments and we need an MLIR-side patch to make sure that the bf16 `math.exp` and `math.log` lower to their LLVM intrinsics

https://github.com/llvm/llvm-project/pull/102971