Assessing the Limits of Program-Specific Garbage Collection Performance
We consider the ultimate limits of program-specific garbage collector performance for real programs. We first characterize the GC schedule optimization problem using Markov Decision Processes (MDPs). Based on this characterization, we develop a method of determining, for a given program run and heap size, an optimal schedule of collections for a non-generational collector. We further explore the limits of performance of a generational collector, where it is not feasible to search the space of schedules to prove optimality. Still, we show significant improvements with Least Squares Policy Iteration, a reinforcement learning technique for solving MDPs. We demonstrate that there is considerable promise to reduce garbage collection costs by developing program-specific collection policies.
Fri 17 JunDisplayed time zone: Tijuana, Baja California change
09:00 - 10:00
|Idle Time Garbage Collection Scheduling|
Ulan Degenbaev Google, Jochen Eisinger Google, Manfred Ernst Google, Ross McIlroy Google, Hannes Payer GoogleMedia Attached
|Assessing the Limits of Program-Specific Garbage Collection Performance|
Nicholas Jacek UMass Amherst, Meng-Chieh Chiu UMass Amherst, Benjamin Marlin UMass Amherst, Eliot Moss University of Massachusetts AmherstMedia Attached