[PATCH] D136677: [AMDGPU] Small cleanups in wait counter code

[Jay Foad via Phabricator via llvm-commits]

foad added a comment.

Note that getScoreLB should //always// be <= getScoreUB, so code like `getScoreLB < getScoreUB` can be updated to `getScoreRange != 0`.



================
Comment at: llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp:1335
 
+static bool isStoreCountWait(const MachineInstr &I) {
+  return (I.getOpcode() == AMDGPU::S_WAITCNT_VSCNT &&
----------------
Not sure about the name. Maybe `isWaitStoreCountZero` would be better...?


================
Comment at: llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp:1336
+static bool isStoreCountWait(const MachineInstr &I) {
+  return (I.getOpcode() == AMDGPU::S_WAITCNT_VSCNT &&
+          I.getOperand(0).getReg() == AMDGPU::SGPR_NULL &&
----------------
Don't need the outermost parens.


================
Comment at: llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp:698
     unsigned LB = getScoreLB(T);
     unsigned UB = getScoreUB(T);
 
----------------
Can use getScoreRange here.


================
Comment at: llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp:759
   const unsigned LB = getScoreLB(T);
   const unsigned UB = getScoreUB(T);
 
----------------
Can use getScoreRange here.


================
Comment at: llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp:1204
     if (ScoreBrackets.getScoreLB(LGKM_CNT) <
             ScoreBrackets.getScoreUB(LGKM_CNT) &&
         ScoreBrackets.hasPendingEvent(SMEM_ACCESS)) {
----------------
Can use getScoreRange here.


================
Comment at: llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp:1598
             ScoreBrackets.getScoreLB(LGKM_CNT) <
                 ScoreBrackets.getScoreUB(LGKM_CNT) &&
             ScoreBrackets.hasPendingEvent(SMEM_ACCESS)) {
----------------
Can use getScoreRange here.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D136677/new/

https://reviews.llvm.org/D136677