salle A. Turing CE4
18 juillet 2013 - 14h00
A Study of the Scalability of Stop-the-world Garbage Collectors on Multicores
par Gaël Thomas de Paris VI / LIP6
Abstract: Large-scale multicore architectures create new challenges for garbage
collectors (GCs). In particular, throughput-oriented stop-the-world
algorithms demonstrate good performance with a small number of cores,
but have been shown to degrade badly beyond approximately 8 cores on a
48-core with OpenJDK 7. This negative result raises the question
whether the stop-the-world design has intrinsic limitations that would
require a radically different approach. Our study suggests that the
answer is no, and that there is no compelling scalability reason to
discard the existing highly-optimised throughput-oriented GC code on
contemporary hardware. We study the default throughput-oriented
garbage collector of OpenJDK 7, called Parallel Scavenge. We identify
its bottlenecks, and show how to eliminate them using well-established
parallel programming techniques. On the SPECjbb2005, SPECjvm2008 and
DaCapo 9.12 benchmarks, the improved GC matches the performance of
Parallel Scavenge at low core count, but scales well, up to 48 cores.
Gaël Thomas is an associate professor at UPMC in Paris, France, where
he joined the faculty in 2006. He is a member of the Regal team, a
joint research team of Inria and UPMC that investigates large-scale
distributed systems. His main research interests include operating
systems, managed runtime environments, large-scale distributed
massively multiplayer games, and multicore programming. He received
the PhD degree from UPMC in 2005 and subsequently performed
postdoctoral research as a member of the Adele Team at the Université
de Grenoble “Joseph Fourier”. Since 2011, he is the chair of the
French chapter of the ACM SIGOPS.
