Reference:

Kiyoshi Kurihara, David Chaiken, and Anant Agarwal. Latency Tolerance through Multithreading in Large-Scale Multiprocessors. Proceedings International Symposium on Shared Memory Multiprocessing, pages 91-101, April 1991.
(pdf, compressed postscript)

Abstract:

In large-scale distributed-memory multiprocessors, remote memory accesses suffer significant latencies. Caches help alleviate the memory latency problem by maintaining local copies of frequently used data. However, they cannot eliminate the latency caused by first-time references and invalidations needed to enforce cache coherence. Multithreaded processors tolerate such latencies by rapidly switching between threads when they encounter cache misses. This paper evaluates the effectiveness of multithreading in Alewife, a scalable multiprocessor that is being developed at MIT. For the applications used in this study, multithreading results in a modest 20% improvement in execution time on a 64-processor machine. The impact of multithreading is expected to be far more significant in larger machines, when remote memory latency becomes a dominant term in the performance equation.

Back to Alewife Publications.
Back to Alewife.

webmaster@cag.lcs.mit.edu $Date: 1998/01/06 16:49:48 $