[llvm] [AMDGPU] Add MaxMemoryClauseSchedStrategy (PR #114957)

[Matt Arsenault via llvm-commits]

================
@@ -561,6 +561,18 @@ createGCNMaxILPMachineScheduler(MachineSchedContext *C) {
   return DAG;
 }
 
+static ScheduleDAGInstrs *
+createGCNMaxMemoryClauseMachineScheduler(MachineSchedContext *C) {
+  const GCNSubtarget &ST = C->MF->getSubtarget<GCNSubtarget>();
+  ScheduleDAGMILive *DAG = new GCNScheduleDAGMILive(
+      C, std::make_unique<GCNMaxMemoryClauseSchedStrategy>(C));
+  DAG->addMutation(createLoadClusterDAGMutation(DAG->TII, DAG->TRI));
+  if (ST.shouldClusterStores())
+    DAG->addMutation(createStoreClusterDAGMutation(DAG->TII, DAG->TRI));
----------------
arsenm wrote:

Does this end up being redundant with the load and store clustering?

https://github.com/llvm/llvm-project/pull/114957