AMDGPU: Avoid constructing new std::vector in initCandidate

Approximately 30% of the time was spent in the std::vector
constructor. In one testcase this pushes the scheduler to being the
second slowest pass.

I'm not sure I understand why these vector are necessary. The default
scheduler initCandidate seems to use some pre-existing vectors for the
pressure.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@371136 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Target/AMDGPU/GCNSchedStrategy.cpp

@@ -71,8 +71,8 @@ void GCNMaxOccupancySchedStrategy::initCandidate(SchedCandidate &Cand, SUnit *SU
   // the tracker, so we need to pass those function a non-const copy.
   RegPressureTracker &TempTracker = const_cast<RegPressureTracker&>(RPTracker);
 
-  std::vector<unsigned> Pressure;
-  std::vector<unsigned> MaxPressure;
+  Pressure.clear();
+  MaxPressure.clear();
 
   if (AtTop)
     TempTracker.getDownwardPressure(SU->getInstr(), Pressure, MaxPressure);

lib/Target/AMDGPU/GCNSchedStrategy.h

@@ -40,6 +40,9 @@ class GCNMaxOccupancySchedStrategy final : public GenericScheduler {
                      const SIRegisterInfo *SRI,
                      unsigned SGPRPressure, unsigned VGPRPressure);
 
+  std::vector<unsigned> Pressure;
+  std::vector<unsigned> MaxPressure;
+
   unsigned SGPRExcessLimit;
   unsigned VGPRExcessLimit;
   unsigned SGPRCriticalLimit;