• DOI:10.1145/2908080.2908120
• Corpus ID: 5243605

Assessing the limits of program-specific garbage collection performance

@article{Jacek2016AssessingTL,
  title={Assessing the limits of program-specific garbage collection performance},
  author={Nicholas Jacek and Meng-Chieh Chiu and Benjamin M Marlin and Eliot Moss},
  journal={Proceedings of the 37th ACM SIGPLAN Conference on Programming Language Design and Implementation},
  year={2016}
}

• Nicholas Jacek, Meng-Chieh Chiu, +1 author E. Moss
• Published 2 June 2016
• Computer Science
• Proceedings of the 37th ACM SIGPLAN Conference on Programming Language Design and Implementation

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.