The Journal of Supercomputing

, Volume 11, Issue 2, pp 119–136

Shared Memory Programming in Metacomputing Environments: The Global Array Approach

  • Jarek Nieplocha
  • Robert Harrison
Article

DOI: 10.1023/A:1007955822788

Cite this article as:
Nieplocha, J. & Harrison, R. The Journal of Supercomputing (1997) 11: 119. doi:10.1023/A:1007955822788

Abstract

The performance of the Global Array shared-memory nonuniform memory-access programming model is explored in a wide-area-network (WAN) distributed supercomputer environment. The Global Array model is extended by introducing a concept of mirrored arrays that thanks to the caching and user-controlled consistency of the shared data structure scan reduce the application sensitivity to the network latency. Latencies and bandwidths for remote memory access are studied, and the performance of a large application from computational chemistry is evaluated using both fully distributed and also mirrored arrays. Excellent performance can be obtained with mirroring if even modest (0.5 MB/s) network bandwidth is available.

Metacomputingshared-memory programmingNUMA memory architectureglobal arraysdistributed arrays

Copyright information

© Kluwer Academic Publishers 1997

Authors and Affiliations

  • Jarek Nieplocha
    • 1
  • Robert Harrison
    • 1
  1. 1.Pacific Northwest National LaboratoryRichland