[Mlir-commits] [mlir] [mlir][gpu] Add extra value types for gpu::ShuffleOp (PR #104605)

[Jakub Kuderski]

kuhar wrote:

> Re the ROCDL lowering, we could reasonably extend the lowering to create one ore more shuffles - and do some bitcasts/zexts for short types

IMO it's better to have a common emulation path since other targets have their own limitations, similar to the pass I linked above.

https://github.com/llvm/llvm-project/pull/104605