Dynamic prediction of collection yield for managed runtimes.pdf

Dynamic Prediction of Collection Yield for Managed Runtimes Michal Wegiel Computer Science Department University of California, Santa Barbara mwegiel@ Chandra Krintz Computer Science Department University of California, Santa Barbara ckrintz@ Abstract The growth in complexity of modern systems makes it in- creasingly difficult to extract high-performance. The soft- ware stacks for such systems typically consist of multiple layers and include managed runtime environments (MREs). In this paper, we investigate techniques to improve cooper- ation between these layers and the hardware to increase the efficacy of automatic memory management in MREs. General-purpose MREs commonly implement parallel and/or concurrent garbage collection and employ com- paction to eliminate heap fragmentation. Moreover, most systems trigger collection based on the amount of heap a program uses. Our analysis shows that in many cases this strategy leads to ineffective collections that are unable to re- claim sufficient space to justify the incurred cost. To avoid such collections, we exploit the observation that dead objects tend to cluster together and form large, never-referenced, re- gions in the address space that correlate well with virtual pages that have not recently been referenced by the appli- cation. We leverage this correlation to design a new, simple and light-weight, yield predictor that estimates the amount of reclaimable space in the heap using hardware page ref- erence bits. Our predictor allows MREs to avoid low-yield collections and thereby improve resource management. We integrate this predictor into three state-of-the-art par- allel compactors, implemented in the HotSpot JVM, that represent distinct canonical heap layouts. Our empirical evaluation, based on standard Java benchmarks and open- source applications, indicates that inexpensive and accurate yield prediction can improve performance significantly. Categories and Subject Descriptors D.3.4 [Programming Languages]: Pr