How lock contention affects energy use in a CMP server

Authors:
Takeshi Ogasawara;Ken Sakamura
Affiliations:
IBM Tokyo Research Laboratory, Yamato, Japan;the University of Tokyo, Tokyo, Japan
Venue:
Proceedings of the 24th ACM SIGPLAN conference companion on Object oriented programming systems languages and applications
Year:
2009

Citing 3
Cited 0

Thin locks: featherweight synchronization for Java

PLDI '98 Proceedings of the ACM SIGPLAN 1998 conference on Programming language design and implementation
Power-Performance Implications of Thread-level Parallelism on Chip Multiprocessors

ISPASS '05 Proceedings of the IEEE International Symposium on Performance Analysis of Systems and Software, 2005
Chameleon: Application-Level Power Management

IEEE Transactions on Mobile Computing

Quantified Score

Hi-index	0.01

Visualization

Abstract

A server application running on a chip multiprocessor with many hardware threads often requires extra processor speed (and energy) because of lock contention. We show that boosting the processor speed for frequently contended critical sections in a Java VM can save more energy on a DVFS-enabled server.