[PATCH] D58728: [MCA] Highlight kernel bottlenecks in the summary view.

[Roman Lebedev via Phabricator via llvm-commits]

lebedev.ri added inline comments.


================
Comment at: include/llvm/MCA/HWEventListener.h:143
+
+  HWPressureEvent(unsigned reason, ArrayRef<InstRef> Insts, uint64_t Mask = 0)
+      : Reason(reason), AffectedInstructions(Insts), ResourceMask(Mask) { }
----------------
```
HWPressureEvent(GenericReason reason, 
```
?


================
Comment at: include/llvm/MCA/HWEventListener.h:147
+  // Reason for this increase in backend pressure.
+  unsigned Reason;
+
----------------
`GenericReason`?


================
Comment at: lib/MCA/HardwareUnits/Scheduler.cpp:232-233
+uint64_t Scheduler::analyzeResourcePressure(SmallVectorImpl<InstRef> &Insts) {
+  for (const InstRef &IR : ReadySet)
+    Insts.emplace_back(IR);
+  return BusyResourceUnits;
----------------
Isn't `ReadySet` a `std::vector`?
It would be best to do something like `Insts.insert(Insts.end(), ReadySet.begin(), ReadySet.end());`
E.g. because this way you avoid constant reallocations since you tell beforehand the amount of entries you want to insert.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D58728/new/

https://reviews.llvm.org/D58728