[llvm] [AMDGPU][AggressiveInstCombine] Narrow 64 bit math to 32 bit if profitable (PR #130577)

[Yingwei Zheng via llvm-commits]

dtcxzyw wrote:

>  I think this is a characteristic of %x, which we don't know at compilation.

You can use `ComputeNumSignBits/ComputeMaxSignificantBits`.


https://github.com/llvm/llvm-project/pull/130577