Fri 17 Jun 2016 09:30 - 10:00 at Grand Ballroom Santa Ynez - Memory Management Chair(s): Sam Guyer

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 Jun

pldi-2016-papers
09:00 - 10:00: Research Papers - Memory Management at Grand Ballroom Santa Ynez
Chair(s): Sam GuyerTufts University
pldi-2016-papers146614680000009:00 - 09:30
Talk
Media Attached
pldi-2016-papers146614860000009:30 - 10:00
Talk
Nicholas JacekUMass Amherst, Meng-Chieh ChiuUMass Amherst, Benjamin MarlinUMass Amherst, Eliot MossUniversity of Massachusetts Amherst
Media Attached