[llvm] [AMDGPU] Revert "Preliminary patch for divergence driven instruction selection. Operands Folding 1." (PR #71710)

[Jay Foad via llvm-commits]

https://github.com/jayfoad commented:

As far as I can tell from the test diffs:
- there's a problem with agpr folding as noted.
- apart from that, SelectionDAG shows no regressions and some slight improvements.
- GlobalISel shows some regressions as noted.

https://github.com/llvm/llvm-project/pull/71710