Dynamic vs. static quantum-based parallel processor allocation

  • Su-Hui Chiang
  • Mary K. Vernon
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1162)

Abstract

This paper improves upon previous synthetic workload models and compares the performance of dynamic spatial equipartitioning (EQS) and the semi-static quantum-based FB-PWS processor allocation defined in [23], under synthetic workloads that have not previously been considered. These new workloads include realistic repartitioning overheads and job characteristics that are consistent with system measurement, anticipated trends, and experience. The overall conclusion from the results is that the EQS policy is generally superior to the FB-PWS policy even under realistic repartitioning overheads. We find cases where the EQS system saturates earlier than the FB-PWS system, and vice versa. This leads to the definition of a modified EQS policy, called EQS-PWS, which has performance equal to or better than EQS and FB-PWS for all workloads examined in this paper.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    R. H. Arpaci, A. C. Dusseau, A. M. Vahdat, L. T. Liu, T. E. Anderson, D. A. Patterson, The Interactions of Parallel and Sequential Workloads on a Network of Workstations. Proc. 1995 ACM Sigmetrics Joint Intl. Conf. on Measurement and Modeling of Computer Systems, Ottawa, pp. 267–278, May 1995.Google Scholar
  2. [2]
    A. Bricker, M. Litzkow, M. Livny, Condor Technical Summary. Technical Report TR 1069, Computer Sciences Dept., University of Wisconsin, Madison, WI, January 1992.Google Scholar
  3. [3]
    R. Chandra, S. Devine, B. Verghese, A. Gupta, M. Rosenblum, Scheduling and Page Migration for Multiprocessor Compute Servers. Proc. 6th Int'l. Conf. on Architectural Support for Programming Languages and Operating Systems (ASPLOS-VI), San Jose, CA, pp. 12–24, October 1994.Google Scholar
  4. [4]
    S.-H. Chiang, R. K. Mansharamani, M. K. Vernon, Use of Application Characteristics and Limited Preemption for Run-to-Completion Parallel Processor Scheduling Policies. Proc. 1994 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, Nashville, TN, pp. 33–44, June 1994.Google Scholar
  5. [5]
    L. W. Dowdy, On the Partitioning of Multiprocessor Systems. Technical Report, Vanderbilt University, July 1988.Google Scholar
  6. [6]
    G. Edjlali, G. Agrawal, A. Sussman, J. Saltz, Data Parallel Programming in an Adaptive Environment. Proc. 9th Int'l. Parallel Processing Symposium Santa Barbara, CA, April 1995.Google Scholar
  7. [7]
    M. J. Feeley, W. E. Morgan, F. H. Pighin, A. R. Karlin, H. M. Levy, C. A. Thekkath, Implementing Global Memory Management in a Workstation Cluster. Proc. Symp. on Operating Systems Principles, Copper Mountain, CO, pp. 201–212, December, 1995.Google Scholar
  8. [8]
    D. G. Feitelson, B. Nitzberg, Job Characteristics of a Production Parallel Scientific Workload on the NASA Ames iPSC/860. Proc. IPPS '95 Workshop on Job Scheduling Strategies for Parallel Systems, Santa Barbara, CA, pp. 337–360, April 1995.Google Scholar
  9. [9]
    D. Ghosal, G. Serazzi, S. Tripathi, The Processor Working Set and Its Use in Scheduling Multiprocessor Systems. IEEE Trans. on Software Engineering, Vol. 17, No. 5, pp. 443–453, May 1991.CrossRefGoogle Scholar
  10. [10]
    S. Hotovy, Workload Evolution on the Cornell Theory Center IBM SP2. Proc. IPPS '96 Workshop on Job Scheduling Strategies for Parallel Systems, Honolulu, Hawaii, April 1996.Google Scholar
  11. [11]
    N. Islam, A. Prodromidis, M. S. Squillante, Dynamic Partitioning in Different Distributed-Memory Environments. Proc. IPPS '96 Workshop on Job Scheduling Strategies for Parallel Systems, Honolulu, Hawaii, April 1996.Google Scholar
  12. [12]
    L. Kleinrock. Queueing Systems, Vol II: Applications. John Wiley & Sons, 1976.Google Scholar
  13. [13]
    S. T. Leutenegger, M. K. Vernon, The Performance of Multiprogrammed Multiprocessor Scheduling Policies. Proceedings of the ACM SIGMETRICS Conference on Measurement & Modeling of Computer Systems, Boulder, CO, pp. 226–236, May 1990.Google Scholar
  14. [14]
    S. Majumdar, D. L. Eager, R. B. Bunt, Scheduling in Multiprogrammed Parallel Systems. Proc. 1988 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, Santa Fe, NM, pp. 104–113, May 1988.Google Scholar
  15. [15]
    S. Majumdar, D. Eager, and R. Bunt. Characterisation of programs for scheduling in multiprogrammed parallel systems. Performance Evaluation, Vol. 13, pp. 109–130, 1991.CrossRefMathSciNetGoogle Scholar
  16. [16]
    R. Mansharamani. Efficient Analysis of Parallel Processor Scheduling Policies. Ph.D. Thesis, Computer Sciences Dept., University of Wisconsin, Madison, WI, November 1993.Google Scholar
  17. [17]
    R. K. Mansharamani, M. K. Vernon, Properties of the EQS Parallel Processor Allocation Policy. Technical Report #1192, Univ. of Wisconsin — Madison Computer Sciences Dept., November 1993.Google Scholar
  18. [18]
    C. McCann, R. Vaswani, J. Zahorjan, A Dynamic Processor Allocation Policy for Multiprogrammed, Shared Memory Multiprocessors. ACM Transactions on Computer Systems, Vol. 11, No. 2, pp. 146–178, May 1993.CrossRefGoogle Scholar
  19. [19]
    V.Naik, S. Setia, and M. Squillante. Performance Analysis of Job Scheduling Policies in Parallel Supercomputing Environments. Proceedings of Supercomputing'93, November 1993.Google Scholar
  20. [20]
    T. D. Nguyen, R. Vaswani, J. Zahorjan, Using Runtime Measured Workload Characteristics in Parallel Processor Scheduling. Proc. IPPS '96 Workshop on Job Scheduling Strategies for Parallel Systems, Honolulu, Hawaii, April 1996.Google Scholar
  21. [21]
    J. K. Ousterhout, Scheduling Techniques for Concurrent Systems, Proc. 3rd Int'l. Conf. on Distributed Computing Systems, pp. 22–30, October 1982.Google Scholar
  22. [22]
    J. D. Padhye, L. W. Dowdy, Dynamic versus Adaptive Processor Allocation Policies for Message Passing Parallel Computers: An Empirical Comparison. Proc. IPPS '96 Workshop on Job Scheduling Strategies for Parallel Systems, Honolulu, Hawaii, April 1996.Google Scholar
  23. [23]
    E. W. Parsons, K. C. Sevcik, Multiprocessor Scheduling for High-Variability Service Time Distributions. Proc. IPPS '95 Workshop on Job Scheduling Strategies for Parallel Systems, Santa Barbara, CA, pp. 127–145, April 1995.Google Scholar
  24. [24]
    V. G. J. Peris, M. S. Squillante, V. K. Naik, Analysis of the Impact of Memory in Distributed Parallel Processing Systems. Proc. 1994 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, Nashville, TN, pp. 5–18, June 1994.Google Scholar
  25. [25]
    K. C. Sevcik, Characterizations of Parallelism in Applications and Their Use in Scheduling. Proc. 1989 ACM SIGMETRICS/Performance '89 Int'l. Conf. on Measurement and Modeling of Computer Systems, Berkeley, CA, pp. 171–180, May 1989.Google Scholar
  26. [26]
    K. C. Sevcik, Application Scheduling and Processor Allocation in Multiprogrammed Parallel Processing Systems. Performance Evaluation, Vol. 19, No. 2/3, pp. 107–140, March 1994.CrossRefGoogle Scholar
  27. [27]
    A. Tucker, A. Gupta, Process Control and Scheduling Issues for Multiprogrammed Shared-Memory Multiprocessors. Proceedings of the 12th ACM Symposium on Operating System Principles, pp. 159–166, December 1989.Google Scholar
  28. [28]
    C.-S. Wu, Processor Scheduling in Multiprogrammed Shared Memory NUMA Multiprocessors, Master's thesis, University of Toronto, 1993.Google Scholar
  29. [29]
    J. Zahorjan, C. McCann, Processor Scheduling in Shared Memory Multiprocessors. Proc. 1990 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, Boulder, CO, pp. 214–225, May 1990.Google Scholar
  30. [30]
    S. Zhou, J. Wang, X. Zheng, P. Delisle, Utopia: A Load Sharing Facility for Large Heterogeneous Distributed Computing Systems. Technical Report, University of Toronto, 1992.Google Scholar

Copyright information

© Springer-Verlag 1996

Authors and Affiliations

  • Su-Hui Chiang
    • 1
  • Mary K. Vernon
    • 1
  1. 1.Computer Sciences DepartmentUniversity of WisconsinMadisonUSA

Personalised recommendations