Multilevel Parallelization of Unsupervised Learning Algorithms in Pattern Recognition on a Roadrunner Architecture

  • Stefan-Gheorghe Pentiuc
  • Ioan Ungurean
Part of the Studies in Computational Intelligence book series (SCI, volume 382)


The aim of the paper is to present a solution to the NP hard problem of determining a partition of equivalence classes for a finite set of patterns. The system must learn the classification of the weighted patterns without any information about the number of pattern classes, based on a finite set of patterns in a metric pattern space. Because a metric is not suitable in all the cases to build an equivalence relation, an ultrametric is generated from indexed hierarchies. The contributions presented in this paper consists in the proposal of multilevel parallel algorithms for bottom-up hierarchical clustering, and hence for generating ultra-metrics based on the metrics provided by the user. The algorithms were synthesized and optimized for clusters having the Roadrunner architecture (the first supercomputer that breaks 1PFlops barrier [1]).


Main Memory Index Matrix Weighted Pattern Unsupervised Learn Algorithm Cell Broadband Engine 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Barker, K.J., Davis, K., Hoisie, A., Kerbyson, D.J., Lang, M., Pakin, S., Sancho, J.C.: Entering the petaflop era: The architecture and performance of Roadrunner. High Performance Computing, Networking, Storage and Analysis, 1–11 (2008), doi:10.1109/SC.2008.5217926Google Scholar
  2. 2.
    Dhillon, I.S., Guan, Y., Kulis, B.: Kernel k-means: spectral clustering and normalized cuts. In: Proceedings of the Tenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2004), pp. 551–556. ACM, New York (2004), doi:10.1145/1014052.1014118CrossRefGoogle Scholar
  3. 3.
    Milligan, G.W., Isaac, P.D.: The validation of four ultrametric clustering algorithms. Pattern Recognition 12(2), 41–50 (1980); doi:10.1016/0031-3203(80)90001-1, ISSN 0031-3203 Google Scholar
  4. 4.
    Celeux, G., Diday, E., et al.: Classification automatiqur de donees. In: Environement statistique et informatique, Dunod Informatique. Dunod Informatique, Bordas Paris (1989)Google Scholar
  5. 5.
    GuoYan, H., DongMei, Z., Jiadong, R., ChangZhen, H.: A Hierarchical Clustering Algorithm Based on K-Means with Constraints. In: Innovative Computing, Information and Control (ICICIC), pp. 1479–1482 (2009), doi:10.1109/ICICIC.2009.18Google Scholar
  6. 6.
    Jun, Q., Qingshan, J., Fangfei, W., Zhiling, H.: A Hierarchical Clustering Based on Overlap Similarity Measure. Software Engineering, Artificial Intelligence, Networking, and Parallel, Distributed Computing 3, 905–910 (2007) ;doi:10.1109/SNPD.2007.502Google Scholar
  7. 7.
    Sandeep, K.: Practical Computing on the Cell Broadband Engine. Springer, Heidelberg (2009); ISBN: 978-1-4419-0307-5Google Scholar
  8. 8.
    Vogt, J.S., Land, R., Boettiger, H., Krnjajic, Z., Baier, H.: IBM BladeCenter QS22: Design, performance, and utilization in hybrid computing systems. IBM Journal of Research and Development 53(5) (2009); doi:10.1147/JRD.2009.5429069Google Scholar
  9. 9.
    Asuncion, A., Newman, D.J.: UCI Machine Learning Repository. University of California, School of Information and Computer Science, Irvine, CA (2007)Google Scholar
  10. 10.
    Guochun, S., Kindratenko, V., Pratas, F., Trancoso, P., Gschwind, M.: Application Acceleration with the Cell Broadband Engine. Computing in Science & Engineering 12(1), 76–81 (2010); doi:10.1109/MCSE.2010.4CrossRefGoogle Scholar
  11. 11.
    Dursun, H., Barker, K.J., Kerbyson, D.J., Pakin, S.: Application profiling on Cell-based clusters. Parallel & Distributed Processing, 1–8 (2009); doi:10.1109/IPDPS.2009.5161092Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Stefan-Gheorghe Pentiuc
    • 1
  • Ioan Ungurean
    • 1
  1. 1.University “Stefan cel Mare” of SuceavaSuceavaRomania

Personalised recommendations