yetingk wrote: > Which presumably means we can't convert from an all ones mask -> unmasked. I think we can still does this transformation for `viota.m`, since all ones mask does not influence the prefix sum calculation. https://github.com/llvm/llvm-project/pull/71483