[llvm] AMDGPU: Mark v_mov_b64_pseudo as a VOP1 instruction (PR #128677)

[Jay Foad via llvm-commits]

jayfoad wrote:

> > Seems reasonable. Which part of the remat implementation cares about the VOP1 flag?
> 
> canRemat, which I think is just hacking around the default skipping rematerialize due to the exec physreg use

Couldn't it check `isVALU` instead of `isVOP1 || isVOP2 || isVOP3`?

https://github.com/llvm/llvm-project/pull/128677