Abstract
Speculative parallelism refers to searching in parallel for a solution, such as finding a pattern in a data base, where finding the first solution terminates the whole parallel process. Different performance prediction methods are required as compared to traditional parallelism. In this paper we introduce an analytical approach to predict the execution time distribution of data-dependent parallel programs that feature N-ary and binary speculative parallel compositions. The method is based on the use of statistical moments which allows program execution time distribution to be approximated at O(1) solution complexity. Measurement results for synthetic distributions indicate an accuracy that lies in the percent range while for empirical distributions on internet search engines the prediction accuracy is acceptable, provided sufficient workload unimodality.
Chapter PDF
Similar content being viewed by others
Keywords
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
References
Gautama, H., van Gemund, A.J.C.: Static performance prediction of datadependent programs. In: ACM WOSP 2000, ottawa,canada,september 2000, pp. 216–226 (2000)
Gautama, H., van Gemund, A.J.C.: Low-cost performance prediction of datadependent data parallel programs. In: Proc. of MASCOTS 2001, Cincinnati, Ohio, pp. 173–182. IEEE Computer Society Press, Los Alamitos (2001)
Gautama, H., van Gemund, A.J.C.: Performance prediction of data-dependent task parallel programs. In: Proc. of EuroPar 2001, Manchester, United Kingdom, August 2001, pp. 106–116 (2001)
Gautama, H.: A statistical approach to performance prediction of speculative parallel programs. Tech. Rep. PDS-2003-007, Delft University of Technology, Delft, The Netherlands (May 2003)
Gelenbe, E., Montagne, E., Suros, R., Woodside, C.M.: Performance of blockstructured parallel programs. In: Cosnard, M., et al. (eds.) Parallel Algorithms and Architectures, pp. 127–138. North-Holland, Amsterdam (1986)
Gumbel, E.J.: Statistical theory of extreme values (main results). In: Sarhan, A.E., Greenberg, B.G. (eds.) Contributions to Order Statistics, pp. 56–93. John Wiley & Sons, New York (1962)
Kruskal, C.P., Weiss, A.: Allocating independent subtasks on parallel processors. IEEE TSE 11, 1001–1016 (1985)
Lester, B.P.: A system for the speedup of parallel programs. In: Proceedings of the 1986 Intl. Conference on Parallel Processing, Maharishi Intl. U, Fairfield, Iowa, pp. 145–152. IEEE, Los Alamitos (1986)
Liang, D.-R., Tripathi, S.K.: On performance prediction of parallel computations with precedent constraints. IEEE TPDS 11(5), 491–508 (2000)
Madala, S., Sinclair, J.B.: Performance of synchronous parallel algorithms with regular structures. IEEE TPDS 2, 105–116 (1991)
Olsson, D.M., Nelson, L.S.: Nelder-Mead simplex procedure for function minimization. Technometrics 17, 45–51 (1975)
Ramberg, J.S., Tadikamalla, P.R., Dudewicz, E.J., Mykytka, F.M.: A probability distribution and its uses in fitting data. Technometrics 21, 201–214 (1979)
Robinson, J.T.: Some analysis techniques for asynchronous multiprocessor algorithms. IEEE TSE 5, 24–31 (1979)
Sahner, R.A., Trivedi, K.S.: Performance and reliability analysis using directed acyclic graphs. IEEE TSE 13, 1105–1114 (1987)
Schopf, J.M., Berman, F.: Performance prediction in production environments. In: Proceedings of IPPS/SPDP-1998, March 30–April 3, pp. 647–653. IEEE Computer Society, Los Alamitos (1998)
Sötz, F.: A method for performance prediction of parallel programs. In: Burkhart, H. (ed.) CONPAR 1990 and VAPP 1990. LNCS, vol. 457, pp. 98–107. Springer, Heidelberg (1990)
Stuart, A., Ord, J.K.: Kendall’s Advanced Theory of Statistics, 6th edn., vol. 1. Halsted Press, New York (1994)
Trivedi, K.S.: Probability and Statistics with Reliability, Queuing and Computer Science Applications. Prentice-Hall, Englewood Cliffs (1982)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Gautama, H., van Gemund, A.J.C. (2003). Symbolic Performance Prediction of Speculative Parallel Programs. In: Kosch, H., Böszörményi, L., Hellwagner, H. (eds) Euro-Par 2003 Parallel Processing. Euro-Par 2003. Lecture Notes in Computer Science, vol 2790. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-45209-6_16
Download citation
DOI: https://doi.org/10.1007/978-3-540-45209-6_16
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-40788-1
Online ISBN: 978-3-540-45209-6
eBook Packages: Springer Book Archive