• DocumentCode
    2737536
  • Title

    Communication overhead for space science applications on the Beowulf parallel workstation

  • Author

    Sterling, Thomas ; Savarese, Daniel ; Becker, Donald J. ; Fryxell, Bruce ; Olson, Kevin

  • Author_Institution
    Center of Excellence in Space Data & Inf. Sci., NASA Goddard Space Flight Center, Greenbelt, MD, USA
  • fYear
    1995
  • fDate
    2-4 Aug 1995
  • Firstpage
    23
  • Lastpage
    30
  • Abstract
    The Beowulf parallel workstation combines 16 PC-compatible processing subsystems and disk drives using dual Ethernet networks to provide a single-user environment with 1 Gops peak performance, half a Gbyte of disk storage, and up to 8 times the disk I/O bandwidth of conventional workstations. The Beowulf architecture establishes a new operating point in price-performance for single-user environments requiring high disk capacity and bandwidth. The Beowulf research project is investigating the feasibility of exploiting mass market commodity computing elements in support of Earth and space science requirements for large data-set browsing and visualization, simulation of natural physical processes, and assimilation of remote sensing data. This paper reports the findings from a series of experiments for characterizing the Beowulf dual channel communication over-head. It is shown that dual networks can sustain 70% greater throughput than a single network alone but that bandwidth achieved is more highly sensitive to message size than to the number of messages at peak demand. While overhead is shown to be high for global synchronization, its overall impact on scalability of real world applications for computational fluid dynamics and N-body gravitational simulation is shown to be modest
  • Keywords
    parallel machines; workstations; Beowulf architecture; Beowulf parallel workstation; Earth science; N-body gravitational simulation; PC-compatible processing subsystems; browsing; communication overhead; computational fluid dynamics; disk drives; dual Ethernet networks; global synchronization; natural physical processes; remote sensing data; scalability; single-user environment; space science; space science applications; visualization; Bandwidth; Computational modeling; Computer architecture; Data visualization; Disk drives; Ethernet networks; Geoscience; Physics computing; Remote sensing; Workstations;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    High Performance Distributed Computing, 1995., Proceedings of the Fourth IEEE International Symposium on
  • Conference_Location
    Washington, DC
  • ISSN
    1082-8907
  • Print_ISBN
    0-8186-7088-6
  • Type

    conf

  • DOI
    10.1109/HPDC.1995.518691
  • Filename
    518691