[Mlir-commits] [mlir] [mlir][scf] Clamp normalized loop sizes where they are combined (PR #223256)

Aviad Cohen llvmlistbot at llvm.org
Wed Sep 23 00:06:28 PDT 2026


https://github.com/AviadCo commented:

I am not sure the fix is correct, Please see my comments.
I suggest to block the transformation in such case instead as this is not fully safe or maintain the signess of the loops in another way.

https://github.com/llvm/llvm-project/pull/223256


More information about the Mlir-commits mailing list