Proceedings Fourteenth Workshop on Parallel and Distributed Simulation (2000)
May 28, 2000 to May 31, 2000
Richard A. Meyer , University of California at Los Angeles
Jay M. Martin , University of California at Los Angeles
Rajive L. Bagrodia , University of California at Los Angeles
Rapid progress in the design of fast CPU chips has outstripped progress in memory and cache performance. Optimistic algorithms would seem to be more vulnerable to poor memory performance because they require extra memory for state saving and anti-messages. We examine the performance of both optimistic and conservative protocols in controlled experiments to evaluate the effects of memory speed and cache size, using a variety of applications.
conservative, optimistic, parallel discrete event simulation, parallel architectures, memory performance, cache performance
R. L. Bagrodia, R. A. Meyer and J. M. Martin, "Slow Memory: The Rising Cost of Optimism," Proceedings Fourteenth Workshop on Parallel and Distributed Simulation(PADS), Bologna, Italy, 2000, pp. 45.