[llvm-branch-commits] [llvm] [AMDGPU] Fold 24 bit multiply with zero low bits (PR #224537)

[Matt Arsenault via llvm-branch-commits]

================

----------------
arsenm wrote:

e.g., you should be able to use Demanded here instead of doing almost this same analysis again 

https://github.com/llvm/llvm-project/pull/224537