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

[Austin Kerbow via llvm-branch-commits]

================
@@ -1018,6 +1018,10 @@ class SIInstrInfo final : public AMDGPUGenInstrInfo {
 
   bool isXDL(const MachineInstr &MI) const;
 
+  /// On gfx1250, XDL VA_VDST decrements are ordered after earlier decrements
+  /// when the destination spans at least eight sequential VGPRs.
----------------
kerbowa wrote:

Against all other VALU instructions of any type.

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