diff options
| author | 2017-10-05 14:35:55 +0100 | |
|---|---|---|
| committer | 2017-10-09 10:39:22 +0100 | |
| commit | e764d2e50c544c2cb98ee61a15d613161ac6bd17 (patch) | |
| tree | 112aa7ca459d2edb4f800897060a2407fcc622c7 /compiler/optimizing/scheduler_arm64.h | |
| parent | ca6fff898afcb62491458ae8bcd428bfb3043da1 (diff) | |
Use ScopedArenaAllocator for register allocation.
Memory needed to compile the two most expensive methods for
aosp_angler-userdebug boot image:
BatteryStats.dumpCheckinLocked() : 25.1MiB -> 21.1MiB
BatteryStats.dumpLocked(): 49.6MiB -> 42.0MiB
This is because all the memory previously used by Scheduler
is reused by the register allocator; the register allocator
has a higher peak usage of the ArenaStack.
And continue the "arena"->"allocator" renaming.
Test: m test-art-host-gtest
Test: testrunner.py --host
Bug: 64312607
Change-Id: Idfd79a9901552b5147ec0bf591cb38120de86b01
Diffstat (limited to 'compiler/optimizing/scheduler_arm64.h')
| -rw-r--r-- | compiler/optimizing/scheduler_arm64.h | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/compiler/optimizing/scheduler_arm64.h b/compiler/optimizing/scheduler_arm64.h index 6682c66ea6..32f161f26a 100644 --- a/compiler/optimizing/scheduler_arm64.h +++ b/compiler/optimizing/scheduler_arm64.h @@ -131,8 +131,8 @@ class SchedulingLatencyVisitorARM64 : public SchedulingLatencyVisitor { class HSchedulerARM64 : public HScheduler { public: - HSchedulerARM64(ScopedArenaAllocator* arena, SchedulingNodeSelector* selector) - : HScheduler(arena, &arm64_latency_visitor_, selector) {} + HSchedulerARM64(ScopedArenaAllocator* allocator, SchedulingNodeSelector* selector) + : HScheduler(allocator, &arm64_latency_visitor_, selector) {} ~HSchedulerARM64() OVERRIDE {} bool IsSchedulable(const HInstruction* instruction) const OVERRIDE { |