RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-26 12:43:36 +01:00
Code Issues Projects Releases Wiki Activity

AMDGPU/SI: Enable testing several variants for si scheduler

Browse Source 
Enable testing different scheduling variants if sgpr usage
is very high. It was previously disabled because of a bug
in handleMove, but it has been fixed since.

Patch by Axel Davy

llvm-svn: 274372
This commit is contained in:
Matt Arsenault 2016-07-01 18:03:46 +00:00
parent b8ff30f5af
commit 792d48b8f4
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
lib/Target/AMDGPU/SIMachineScheduler.cpp
View File

@ -1826,7 +1826,7 @@ void SIScheduleDAGMI::schedule()
SIScheduler Scheduler(this);
Best = Scheduler.scheduleVariant(SISchedulerBlockCreatorVariant::LatenciesAlone,
SISchedulerBlockSchedulerVariant::BlockLatencyRegUsage);
#if 0 // To enable when handleMove fix lands
// if VGPR usage is extremely high, try other good performing variants
// which could lead to lower VGPR usage
if (Best.MaxVGPRUsage > 180) {
@ -1865,7 +1865,7 @@ void SIScheduleDAGMI::schedule()
Best = Temp;
}
}
#endif
ScheduledSUnits = Best.SUs;
ScheduledSUnitsInv.resize(SUnits.size());