[Mlir-commits] [mlir] [ROCDL] Use attached target on the GPU module when lowering GPU ops to ROCDL (PR #110735)

[Fabian Mora]

fabianmcg wrote:

> The one thing I'd like to get a touch of consensus on is whether this will help with solving the target information issues in the long run (ex. letting something like `--convert-to-llvm` pick up the data layout)

@krzysz00 not really, the infer mechanism  is not compatible with with how `convert-to-llvm` works.

https://github.com/llvm/llvm-project/pull/110735