Connectivity-based garbage collection

Abstract

We introduce a new family of connectivity-based garbage collectors (Cbgc) that are based on potential objectconnectivity properties. The key feature of these collectors is that the placement of objects into partitions is determined by performing one of several forms of connectivity analyses on the program. This enables partial garbage collections, as in generational collectors, but without the need for any write barrier.The contributions of this paper are 1) a novel family of garbage collection algorithms based on object connectivity; 2) a detailed description of an instance of this family; and 3) an empirical evaluation of Cbgc using simulations. Simulations help explore a broad range of possibilities for Cbgc, ranging from simplistic ones that determine connectivity based on type information to oracular ones that use run-time information to determine connectivity. Our experiments with the oracular Cbgc configurations give an indication of the potential for Cbgc and also identify weaknesses in the realistic configurations. We found that even the simplistic implementations beat state-of-the-art generational collectors with respect to some metrics (pause times and memory footprint).

A Phase-Adaptive Garbage Collector Using Dynamic Heap Partitioning and Opportunistic Collection

“…We also improve GC efficiency during the ramp phase. ConnectivityBased GC (CBGC) [16] uses the connectivity (gathered by pointer analysis) to statically partition the heap and to determine which region to collect. Like our collector, it focuses GC effort on the regions yielding the good GC efficiency.…”

“…Our phase-adaptive garbage collector (hereafter abbreviated as PAGC) exploits application phases 1) to group objects according to the similarities of their lifetime [15], [16] and 2) to determine the best GC time for each group [18]- [20]. PAGC consists of the components listed below:…”

A Fast Analysis for Thread-Local Garbage Collection with Dynamic Class Loading

“…Our figures for analysis time and space show a 100x and a 20x improvement over the only other analysis of which we are aware that supports dynamic class loading [18]. However, their results were obtained from a 2.4GHz Pentium 4 with 2GB memory running Linux, kernel 2.4.…”

“…• It uses less time and space than other analyses that accommodate dynamic class loading [18]. It is sufficiently fast to make incorporation into a production JVM (Sun's ExactVM for Solaris) realistic.…”

“…Hirzel et al [18] describe a Anderson [4] pointer analysis that supports all Java features including dynamic class loading. The memory and runtime costs of their analysis are significantly larger than ours, although comparisons between our JVMs are hard to draw.…”