[llvm] [AMDGPU] Constant folding for wave-reduce intrinsics (PR #212755)

[via llvm-commits]

easyonaadit wrote:

> This should be a false positive. IIRC these intrinsics only accept scalar inputs, right?

Yupp, currently they only work for scalar inputs. But good catch, I think there should be a more graceful way to handle this failure. Will tackle it in a follow up.



https://github.com/llvm/llvm-project/pull/212755