[llvm] [AMDGPU] Introduce a pseudo mnemonic for S_DELAY_ALU in MIR. (PR #96004)

[Jay Foad via llvm-commits]

================
@@ -0,0 +1,192 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
+# RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=-wavefrontsize32,+wavefrontsize64 -verify-machineinstrs -run-pass=amdgpu-insert-delay-alu %s -o - | FileCheck %s
----------------
jayfoad wrote:

If you don't you end up with both +wavefrontsize32 and +wavefrontsize64 set which can lead to weird inconsistencies. Since #86957 we check for the inconsistency, but only during isel so the check never runs on MIR test cases. I'd love to hear ideas for better places to put the check.

https://github.com/llvm/llvm-project/pull/96004