An extended gradient model for NUMA multiprocessor systems

  • Feixiong Liu
  • Thomas Peikenkamp
  • Werner Damm
Concurrency and Networking
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1023)


In this paper, we present the design and implementation of an effective and scalable dynamic load balancing system for Non-Uniform Memory Access (NUMA) multiprocessors where load balancing is a key issue to achieve adequate efficiency. The proposed load balancing algorithm extends the well-known gradient model to enhance its applicability in a wide range of multiprocessor systems and to improve the overall system performance. A comparative performance study between the two models based on the preliminary simulation results is also reported in the paper.


Load balancing Multiprocessing Gradient Model 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    I. Ahmad and A. Ghafoor, Semi-distributed load balancing for massively parallel multicomputer systems, IEEE trans. Software Eng. 17(10) (Oct. 1991) 987–1004.Google Scholar
  2. 2.
    K.A.M. Ali and M. Wang, An investigation of an OR-parallel model for Horn clause programs, SICS research report, Sept, 1986.Google Scholar
  3. 3.
    M.J. Berger adn S. Bokhari, A partitioning strategy for non-uniform problems on multiprocessors, IEEE Trans. Computer, C-36, 5 (1987), 570–580.Google Scholar
  4. 4.
    T. Casavant and J.G. Kuhl, A taxonomy of scheduling in general-purpose distributing systems, IEEE Trans. Software Eng., 14(2) (Feb. 1988) 141–154.Google Scholar
  5. 5.
    R. Jain, The art of computer systems performance analysis, John Wiley and Sons, 1991.Google Scholar
  6. 6.
    F. Lin and R. Keller, The gradient model load balancing method, IEEE Trans. Software Eng., 13(1), (January 1987), 32–38.Google Scholar
  7. 7.
    R. Lulin, B. Monien and F. Ramme, Load balancing in large networks: a comparative study, 3rd IEEE Symp. on Parallel and Distributed Processing, Dallas 1991, 329–336.Google Scholar
  8. 8.
    E. Mohr, D. A. Kranz and R. H. Halstead, Lazy Task Creation, A technique for Increasing the Granularity of Parallel Programs, IEEE Transactions on Parallel and distributed systems, 1990.Google Scholar
  9. 9.
    F.J. Muniz and E.J. Zaluska, Parallel Load-balancing: An extension to the gradient model, Parallel Computing, No. 21, Jan, 1995.Google Scholar
  10. 10.
    N. G. Shivaratri, P. Krueger, and M. Singhal, Load distributing for locally distributed systems, IEEE Computer, December, 1992.Google Scholar
  11. 11.
    S. Nishimura and T.L. Kunii, A decentralized dynamic scheduling scheme for transputer networks, in T.L. Kunii and D, Nay, eds., Proc. 3rd Transputer/OCCAM int. conf., May, 1990, Tokyo, Japan.Google Scholar
  12. 12.
    R.D. Rettberg, W.R. Crowther, R.P. Carvey and R.S. Tomlinson, The monarch parallel processor hardware design, IEEE computer, April, 1990, 18–30.Google Scholar
  13. 13.
    V. Sarkar and J. Hennesey, Compile-time partitioning and scheduling of parallel programs, Proc. SIGPLAN'86 symp. compiler construction, 21(7), 1986, 17–26.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1995

Authors and Affiliations

  • Feixiong Liu
    • 1
  • Thomas Peikenkamp
    • 1
  • Werner Damm
    • 1
  1. 1.FB InformatikOldenburg UniversitätOldenburgGermany

Personalised recommendations