[llvm] [AMDGPU][VOPD] Limit VOPDPairing from reaching over load dependencies (PR #201930)

[Jay Foad via llvm-commits]

jayfoad wrote:

> Mutations are applied before the schedule is decided

Right, and I think that makes the whole concept of this patch pretty dubious. Wouldn't it be better if we could still add the VOPD pairing edge, but somehow tell the scheduler to treat it with lower priority than memory clustering edges?

https://github.com/llvm/llvm-project/pull/201930