[llvm] [X86] Shrink i32 unsigned compares to i8/i16 when high bits are zero (PR #197801)

[Weiwen He via llvm-commits]

he-weiwen wrote:

> Make sure tests are added as the first commit (with baseline codegen) in the PR, so then later commits in the chain show the codegen diffs - this all then squashes when we push to main

fixed

https://github.com/llvm/llvm-project/pull/197801