Quantifying and Improving the Performance of Garbage Collection

Researchers have devoted decades towards improving garbage collection performance.A key question is: what opportunity remains for performance improvement? Because theperceived performance disadvantage of garbage collection is often cited as a barrier to itsadoption, we propose using explicit memory management as the touchstone for evaluatinggarbage collection performance.This work begins by asking how well garbage collection performs currently and which areas remain for it to improve. We present oracular memory management to help answer this question by enabling us to quantify garbage collection?s performance. Our results show that while garbage collection can perform competitively, this good performance assumesthe heap fits in available memory. Our work thus focuses onto this brittleness, and we show the designand implement of our bookmarking collector. While performing competitively when notpaging, our new collector?s paging performance helps make garbage collection?s performancemore robust.

Autorentext
Hertz, Matthew Matthew Hertz is an Assistant Professor of Computer Science at Canisius College. His papers have been published in OOPSLA, PLDI, TOPLAS, and other top venues. He lives in Buffalo, NY with his wife, Sarah, and daughter, Shoshanna. More of his writing can be found at http://cs.canisius.edu/~hertzm.