Sage Journals: Discover world-class research

Abstract

This simulation study investigated the relationship between workload characteris tics and process speedup obtainable on a hypercube parallel processor architecture. There were two goals: the first was to determine the functionat relationship between workload characteristics and speedup, and the second was to show how simulation could be used to model the concurrently executing processes to allow estimation of such a relation. The hypercube implementation used in this study was a packet-switched network with predeter mined routing and a balanced computa tional workload. Three independent vari ables were controlled: total computational workload, number of processors and the message traffic load. A benchmark program was used to estimate the fundamental timing models and to validate a discrete event simulation. Results of this study are useful to software designers seeking to predict the degree of performance improve ment attainable on a hypercube class machine. The methodology and results can be extended to other parallel processing architectures.

Get full access to this article

View all access options for this article.

References

[FE85] Felten, Edward, et al. , "The Traveling Salesman Problem on a Hypercube, MIMD Computer," Proceedings of the 1985 International Conference on Parallel Processing. 6-10 Washington, DC: IEEE Computer Society Press, 1985.

[GU87] Gutzmann, Kurt M. , "Optimal Dimension of Hypercubes for Sorting," Computer Architecture News, 15: 68-72 (March 1987 ).

[HW84] Hwang, Kai and Faye A. Briggs. Computer Architecture and Parallel Processing New York: McGraw Hill Book Company, 1984.

[IN86a] Indurkhya, Bipin et al.

"Optimal Partitioning of Randomly Generated Distribution Programs,"

IEEE Transactions on Software Engineering , SE-12: 483-495 (March 1986).

[IN86b] Intel Corporation. iPSC Software Internal Product Specification Beaverton, OR, April 1986.

[LE86] LeBlanc, Thomas J. "Shared Memory Versus Message Passing in a Tightly-Coupled Multiprocessor: A Case Study," Proceedings of the 1986 International Conference on Parallel Processing. 463-466 New York: IEEE Press, 1986.

[MO87] Moore, Timothy S. "A Simulation Study of a Parallel Processor with Unbalanced Loads." MS thesis, AFIT/GSC/ENG/ 87D-20 Air Force Institute of Technology (AU), Wright Patterson Air Force Base, OH, December 1987.

[M87a] Ni, Lionel M., et al. "Parallel Algorithm Design Considerations for Hypercube Mulitprocessors," Proceedings of the 1987 International Conference on Parallel Processing. 717-720 University Park, PA: The Pennsylvania State University Press, 1987 .

[NI87b] Nicol, David M. and Frank H. Willard. "Problem Size, Parallel Architecture, and Optimal Speedup," Proceedings of the 1987 International Conference on Parallel Processing. 347-354 University Park, PA: The Pennsylvania State University Press, 1987.

10.

[PR86] Pritsker, A. Alan B. Introduction to Simulation and SLAM II New York: John Wiley and Sons, 1986.

11.

[RE88] Reed, Daniel A. , Assistant Professor. Telephone interview Department of Computer Science, University of Illinois at Urbana-Champaign, 3 May 1988.

12.

[RE87] Reed, Daniel A. and Dirk C.

Grunwald. "The Performance of Multicomputer Interconnection Networks,"

Computer, 20: 63-73 (June 1987 ).

13.

[ST87] Stone, Harold S. High Performance Computer Architecture Reading, MA: Addison-Wesley Publishing Company, 1987.

14.

[SE85] Seitz, Charles L.

"The Cosmic Cube,"

Communications of the ACM , 28: 22-33 (January 1985).

15.

[WI87] Wiley, Paul.

"A Parallel Architecture Comes of Age at Last,"

IEEE Spectrum , 24: 46-50 (June 1987 ).

A performance study of the hypercube parallel processor architecture

Abstract

Get full access to this article

References