Codegen for switch of `llvm.ucmp` could be optimised

> https://godbolt.org/z/6dP4je7Eb
>
> Looks like LLVM's middle-end optimizes that perfectly well -- it gives exactly the `llvm.ucmp.i8.i8` intrinsic and the `switch` as you'd expect -- but LLVM's x64 backend doesn't optimize it as well as it could.

_Originally posted by @scottmcm in [#5354](https://github.com/rust-lang/rust-clippy/issues/5354#issuecomment-3761928995)_

Start with the Godbolt example at https://godbolt.org/z/6dP4je7Eb and compare the generated x64 code for the llvm.ucmp intrinsic and switch. Read the LLVM x64 backend codegen path involved in this case; done means the switch is lowered to the more efficient code expected from the example.

Written by the indexing model from the issue text.