[Mlir-commits] [mlir] Add Lowerings for GPU WMMA F16/F32 ops to ROCDL dialect (PR #69357)

[Navdeep Katel]

navdeepkk-polymagelabs wrote:

Hi @krzysz00, can we take the two phase lowering as the next step (in a subsequent PR) and land this one first (as this is already functional and covers a good amount of cases for RDNA 3)? I can fix the testing related issues though so they are robust.

https://github.com/llvm/llvm-project/pull/69357