[llvm] [X86][AVX] Prefer VPSRAV to VPSRA style shifts for known splats #39424 (PR #87913)

[Simon Pilgrim via llvm-commits]

RKSimon wrote:

Also, we need to update the PR title - the original issue title was over specific - "[X86][AVX] Prefer per-element vector shifts for known splats"

https://github.com/llvm/llvm-project/pull/87913