Neural network classifiers execution on superscalar microprocessors

  • Omar Hammami
II System Architecture
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1615)


This paper evaluates the contribution of various microprocessor architectural features on the execution of 4 neural networks used for classification problems. In this study, we selected the grnn, pnn, mnn and rbfn networks trained for the Iris data set and simulated with 10,000 elements datasets. Using a superscalar simulator we evaluated various architectural parameters such as IPC, memory hierarchy, branch prediction, functionnal units configuration. The main contribution of this work is to show that neural network workloads deserve their own characterization which cannot be derived from SPEC95 characteristics.

Key words

classification microarchitecture neural network performance superscalar 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    S. Kung and J. Hwang, “Neural Networks for Intelligent Multimedia Processing”, Proceedings of the IEEE, Vol. 86, No. 6, June 1988.Google Scholar
  2. 2.
    T. Cooley and E. Micheli Tzanakou, “Classification of Mammograms Using a Modular Neural Network”, Journal of Intelligent Systems, Vol. 8, Nos 1–2, 1998.Google Scholar
  3. 3.
    SPEC Newsletter, Sept. 1995. Information about the SPEC 95 benchmarks used in this study is available at: Scholar
  4. 4.
    R. Weicker, “On the Use of SPEC Benchmarks in Computer Architecture Research”, Computer Architecture News, pp. 19–22, Vol. 25, No. 1, March 1997.CrossRefGoogle Scholar
  5. 5.
    D.C. Lee, P.J. Crowley and J.L. Baer, “Execution Characteristics of Desktop Applications on Windows NT”, in Proc. of ISCA 25, pp. 27–38, Barcelona, Spain, June 27–July 1st, 1998. (NT Traces available at Google Scholar
  6. 6.
    K. Keeton, D. Patterson, Y. He, Re. Raphael and W. Baker, “Performance Characterization of a Quad Pentium Pro SMP using OLTP Workloads”, in Proc. of ISCA 25, pp. 15–26, Barcelona, Spain, June 27–July 1st, 1998.Google Scholar
  7. 7.
    D Bhandarkar and J. Ding, “Performance Characterization of the Pentium Pro Processor”, in Proc. 3rd HPCA, Deb. 1–5, 1997. (available at Google Scholar
  8. 8.
    R. Jain, “The Art of Computer Systems Performance Analysis”, John Wiley and Sons, Inc., 1991.Google Scholar
  9. 9.
    R.H. Saavedra and A.J. Smith, “Analysis of Benchmark Characteristics and Benchmark Performance Prediction,” ACM Transactions on Computer Systems, pp. 344–384, Vol. 14, No. 4, Nov. 1996.CrossRefGoogle Scholar
  10. 10.
    D.F. Specht, “A General Regression Neural Network”, IEEE Trans. on Neural Networks, Vol. 2, No. 6, pp. 568–576, 1991.CrossRefGoogle Scholar
  11. 11.
    R.A. Jacobs, M.I. Jordan, S.J. Nowlan and G.E. Hinton, “Adaptive Mixtures of Local Experts”, Neural Computation, vol. 3, pp. 79–87, 1991.Google Scholar
  12. 12.
    D.F. Specht, “Probabilistic Neural Networks”, IEEE Trans. on Neural Networks, Vol. 2, No. 6, pp. 568–576, 1991.CrossRefGoogle Scholar
  13. 13.
    J. Moody and C.J. Darken, “Fast learning in Networks of Locally Tuned Processing Units”, Neural Computation, 1, pp. 281–294, 1989.Google Scholar
  14. 14.
    Neural Works Professionnal II Plus, Neural Ware Inc, 1995.Google Scholar
  15. 15.
    D. Burger and T. Austin, “The SimpleScalar Tool Set, version 2.0”, University of Wisconsin Madison, Computer Science Department, Technical report 1342, June 1997.Google Scholar
  16. 16.
    B. Black and J.P. Shen, “Calibration of Microprocessor Performance Models”, IEEE Computer, pp. 59–65, May 1998.Google Scholar
  17. 17.
    M. Reilly and J. Edmondson, “Performance Simulation of an Alpha Microprocessor”, IEEE Computer, pp. 50–58, May 1998.Google Scholar
  18. 18.
    M. Johnson, “Superscalar Microprocessor Design”, Prentice Hall Series in Innovative technology, Prentice Hall Inc, 1991.Google Scholar
  19. 19.
    D.B. Noonburg and J.P. Shen, A Framework for Statistical Modeling of Superscalar Performance”, in Proc. 3rd HPCA, Deb. 1–5, 1997 (available at Google Scholar
  20. 20.
    Y. Zhang and G.B. Adams III, “Performance Modeling and Code Partitioning for the DS Architecture”, in Proc. of ISCA 25, pp. 293–304, Barcelona, Spain, June 27–July 1st, 1998.Google Scholar
  21. 21.
    Y. Zhu and W.F. Wong, “The Effect of Instruction Dependency on Superscalar Processor Performance” in Proc. ACAC98.Google Scholar
  22. 22.
    M. Matson, D. Bailey, S. Bell, L. Biro, S. Butler, J. Clouser, J. Farrell, M. Gowan, D. Priore, K. Wilcox, “Circuit Implementation of a 600 MHz Superscalar RISC Microprocessor”, in Proc. of ICCD98, Austin, TX, Oct. 5–7, 1998.Google Scholar
  23. 23.
    R.E. Kessler, “The Alpha 21264 Microprocessor: Out-Of-Order Execution at 600 Mhz”, in Proc. of Hot Chips 10, Stanford University, Palo Alto, Aug. 16–18, 1998.Google Scholar
  24. 24.
    K.C. Yeager, “The MIPS R10000 Superscalar Microprocessor”, IEEE Micro, pp. 28–40, April 1996.Google Scholar
  25. 25.
    Intel Corp., “Pentium II Xeon Processor Technology Brief”, Order Number 243792-002, September 1998.Google Scholar
  26. 26.
    Intel Corp., “Pentium II Xeon Processor at 400 and 450 Mhz”, Order Number 243770-002, June 1998.Google Scholar
  27. 27.
    G. Lauterbach, “UltraSPARC-III: A 600 MHz 64-bit Superscalar Processor for 1000-way Scalable Systems”, in Proc. of Hot Chips 10, Stanford University, Palo Alto, Aug. 16–18, 1998.Google Scholar
  28. 28.
    D.A. Patterson and J.L. Hennessy, “Computer Organization and Design”, 2nd Ed., Morgan Kaufman Publishers, Inc., 1998.Google Scholar
  29. 29.
    S. Wallace and N. Bagherzadeh, “Modeled and Measured Instruction fetching Performance for Superscalar Microprocessors”, IEEE Trans. on PDS, Vol. 9, No. 6, pp. 570–578, June 1998.Google Scholar
  30. 30.
    M.J. Charney and T.R. Puzak, “Prefetching and Memory System Behavior of the SPEC 95 benchmark Suite”, IBM J. Res. Develop. Vol. 41, No. 3, pp. 265–286, May 1997.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • Omar Hammami
    • 1
  1. 1.University of AizuFukushimaJapan

Personalised recommendations