[Mlir-commits] [mlir] [MLIR][XeGPU] Lower arith.truncf wider than one xevm.truncf group (PR #217130)

[Artem Kroviakov]

https://github.com/akroviakov commented:

Isn't it the task of the distribution to decompose shapes to match the instruction restrictions?

https://github.com/llvm/llvm-project/pull/217130