Research Article

Performance evaluation of the SX-6 vector architecture for scientific computations

  1. Leonid Oliker1,
  2. Andrew Canning1,
  3. Jonathan Carter1,
  4. John Shalf1,
  5. David Skinner1,
  6. Stéphane Ethier2,
  7. Rupak Biswas3,*,
  8. Jahed Djomehri3,†,
  9. Rob Van der Wijngaart3,‡

Article first published online: 29 NOV 2004

DOI: 10.1002/cpe.884

Issue

Concurrency and Computation: Practice and Experience

Concurrency and Computation: Practice and Experience

Volume 17, Issue 1, pages 69–93, January 2005

Additional Information

How to Cite

Oliker, L., Canning, A., Carter, J., Shalf, J., Skinner, D., Ethier, S., Biswas, R., Djomehri, J. and Van der Wijngaart, R. (2005), Performance evaluation of the SX-6 vector architecture for scientific computations. Concurrency Computat.: Pract. Exper., 17: 69–93. doi: 10.1002/cpe.884

Author Information

  1. 1

    CRD/NERSC, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, U.S.A.

  2. 2

    Princeton Plasma Physics Laboratory, Princeton University, Princeton, NJ 08453, U.S.A.

  3. 3

    NAS Division, NASA Ames Research Center, Moffett Field, CA 94035, U.S.A.

  1. †

    Employee of Computer Sciences Corporation

  2. ‡

    Employee of Computer Sciences Corporation

*NAS Division, NASA Ames Research Center, Moffett Field, CA 94035, U.S.A.

Publication History

  1. Issue published online: 29 NOV 2004
  2. Article first published online: 29 NOV 2004
  3. Manuscript Accepted: 9 MAR 2004
  4. Manuscript Revised: 9 FEB 2004
  5. Manuscript Received: 15 AUG 2003

SEARCH

SEARCH BY CITATION

ARTICLE TOOLS

Get PDF (849K)

Keywords:

  • microbenchmarks;
  • NAS Parallel Benchmarks;
  • scientific applications;
  • vectorization;
  • superscalar performance

Abstract

The growing gap between sustained and peak performance for scientific applications is a well-known problem in high-performance computing. The recent development of parallel vector systems offers the potential to reduce this gap for many computational science codes and deliver a substantial increase in computing capabilities. This paper examines the intranode performance of the NEC SX-6 vector processor, and compares it against the cache-based IBM Power3 and Power4 superscalar architectures, across a number of key scientific computing areas. First, we present the performance of a microbenchmark suite that examines many low-level machine characteristics. Next, we study the behavior of the NAS Parallel Benchmarks. Finally, we evaluate the performance of several scientific computing codes. Overall results demonstrate that the SX-6 achieves high performance on a large fraction of our application suite and often significantly outperforms the cache-based architectures. However, certain classes of applications are not easily amenable to vectorization and would require extensive algorithm and implementation reengineering to utilize the SX-6 effectively. Copyright © 2005 John Wiley & Sons, Ltd.

Get PDF (849K)