[llvm] [X86] Use the standard cmp+cmov for select (X != 0), -1, Y if we will be setting sbb to 0 anyway (PR #149672)

[Simon Pilgrim via llvm-commits]

RKSimon wrote:

I think I would like to see benchmarks that show this to be a better approach.

https://github.com/llvm/llvm-project/pull/149672