[llvm] [AMDGPU][MachineScheduler] Alternative way to control excess RP. (PR #68004)

[Jeffrey Byrnes via llvm-commits]

================
@@ -959,16 +970,6 @@ void GCNSchedStage::checkScheduling() {
                       << DAG.MinOccupancy << ".\n");
   }
 
-  unsigned MaxVGPRs = ST.getMaxNumVGPRs(MF);
-  unsigned MaxSGPRs = ST.getMaxNumSGPRs(MF);
-  if (PressureAfter.getVGPRNum(false) > MaxVGPRs ||
-      PressureAfter.getAGPRNum() > MaxVGPRs ||
-      PressureAfter.getSGPRNum() > MaxSGPRs) {
-    DAG.RescheduleRegions[RegionIdx] = true;
----------------
jrbyrnes wrote:

Why do we drop the maxNumGPR checks -- these are wavesPerEU aware?

https://github.com/llvm/llvm-project/pull/68004