Ocean warning: avoid drowning

ABSTRACT

Ocean is a popular program from the SPLASH-2 parallel benchmark suite. A complete application, as opposed to a computational kernel, Ocean is often used as a representative of a well-tuned parallel program in architectural studies. However, we find there is a danger in using Ocean to evaluate proposed enhancements that purport to either improve scalability or reduce synchronization overhead. The default Ocean code contains an ill-advised code segment that seriously handicaps its "base" performance in common architectural studies. We provide the one-line fix for the offending code that improves performance by as much as a factor of 2.3, and suggest that architecture researchers using Ocean to evaluate their new ideas---especially when discussing scalability or synchronization---change their code immediately.

REFERENCES

• 1: Jeff Gibson , Robert Kunz , David Ofelt , Mark Horowitz , John Hennessy , Mark Heinrich, FLASH vs. (Simulated) FLASH: closing the simulation loop, Proceedings of the ninth international conference on Architectural support for programming languages and operating systems, p.49-58, November 2000, Cambridge, Massachusetts, United States
• 2: James Laudon , Daniel Lenoski, The SGI Origin: a ccNUMA highly scalable server, Proceedings of the 24th annual international symposium on Computer architecture, p.241-251, June 01-04, 1997, Denver, Colorado, United States
• 3: Steven Cameron Woo , Moriyoshi Ohara , Evan Torrie , Jaswinder Pal Singh , Anoop Gupta, The SPLASH-2 programs: characterization and methodological considerations, Proceedings of the 22nd annual international symposium on Computer architecture, p.24-36, June 22-24, 1995, S. Margherita Ligure, Italy