Author:
Horie Takeshi,Hayashi Kenichi,Shimizu Toshiyuki,Ishihata Hiroaki
Abstract
The performance of message-passing applications depends on cpu speed, communication throughput and latency, and message handling overhead. In this paper we investigate the effect of varying these parameters and applying techniques to reduce message handling overhead on the execution efficiency of ten different applications. Using a message level simulator set up for the architecture of the AP1000, we showed that improving communication performance, especially message handling, improves total performance. If a cpu that is 32 times faster is provided, the total performance increases by less than ten times unless message handling overhead is reduced. Overlapping computation with message reception improves performance significantly. We also discuss how to improve the AP1000 architecture.
Publisher
Association for Computing Machinery (ACM)
Reference24 articles.
1. nCUBE$ Supercomputers Technical Overview 1990. nCUBE$ Supercomputers Technical Overview 1990.
2. Interprocessor communication speed and performance in distributed-memory parallel processors
3. R. Adauskas: "iPSC/2 system: a second generation hypercube " In Third Conlerence on Hypercube Concurrent Computers and Applications pp. 38-42 Jan. 1988. 10.1145/62297.62303 R. Adauskas: "iPSC/2 system: a second generation hypercube " In Third Conlerence on Hypercube Concurrent Computers and Applications pp. 38-42 Jan. 1988. 10.1145/62297.62303
4. Improved multithreading techniques for hiding communication latency in multiprocessors
5. R. P. Brent: "The LINPACK benchmark on the AP1000 " In Fourth Symposium on the Frontiers of Massively Parallel Computation Oct. 1992. R. P. Brent: "The LINPACK benchmark on the AP1000 " In Fourth Symposium on the Frontiers of Massively Parallel Computation Oct. 1992.
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献