[llvm] [AMDGPU] Model ordered XDL writes in expert scheduling (PR #218326)

[Jay Foad via llvm-commits]

================
@@ -0,0 +1,445 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 6
----------------
jayfoad wrote:

It would be helpful to precommit this (or use stacked PRs or at least two separate commits in one PR) so that we can see the effect of your patch.

https://github.com/llvm/llvm-project/pull/218326