Regular article
A methodology for performance evaluation of parallel applications on multiprocessors

https://doi.org/10.1016/0743-7315(92)90093-3Get rights and content

Abstract

The potential for high performance exhibited by multiprocessors, due to the available processor parallelism, can only be attained if the application can be adequately decomposed into parallel tasks which cooperate in order to achieve the desired goal. In this paper we present a methology for evaluating the performance of parallel applications that run on shared memory multi-processors. This methodology consists of an efficient analytic technique which allows us to obtain the average execution time of a parallel application taking into account its structure, defined by a task graph, several task parameters, and the contention caused by the processor-memory interconnection network.

References (31)

  • A. Kapelnikov et al.

    A modeling methodology for the analysis of concurrent systems and computations

    J. Parallel Distrib. Comput.

    (1989)
  • L.A. Barroso

    A methodology for the performance evaluation of scientific applications in multiprocessors

  • L.A. Barroso et al.

    Performance of processor-memory interconnection networks under unbalanced traffic

  • E.G. Coffman et al.

    Operating Systems Theory

    (1973)
  • Cray Research, Inc.

    Cray X-MP Series Mainframe Reference Manual

    (1982)
  • G. Dahlquist et al.

    Numerical Methods

    (1974)
  • F. Darema-Rogers

    Parallel applications performance methodology

  • D.M. Dias et al.

    Analysis and simulation of buffered delta networks

    IEEE Trans. Comput.

    (April 1981)
  • J. Dongarra Jack et al.

    Advanced architecture computers

    The University of Tennessee Tech. Rep. CS-89–90

    (Nov. 1989)
  • R. Goke et al.

    Banyan networks for partitioning on multiprocessor systems

  • A. Gottlieb et al.

    The NYU Ultracomputer—Designing a MIMD, shared memory parallel machine

  • V. Guarna et al.

    FAUST: An integrated environment for the development of parallel programs

  • W.D. Hillis et al.

    The Connection Machine

    (1985)
  • K.B. Irani et al.

    Performance study of a clustered shared-memory multiprocessor

  • L. Kleinrock

    • Cited by (0)

    View full text
    Copyright © 1992 Published by Elsevier Inc.