Advertisement

First graph partitioning and its application in sparse matrix ordering

  • Anshul Gupta
  • Fred Gustavson
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1184)

Abstract

Graph partitioning is a fundamental problem in several scientific and engineering applications, including task partitioning for parallel processing. In this paper, we describe heuristics that improve the state-of-the-art practical algorithms used in graph partitioning software in terms of both partitioning speed and quality. An important use of graph partitioning is in ordering sparse matrices for obtaining direct solutions to sparse systems of linear equations arising in engineering and optimization applications. The experiments reported in this paper show that the use of these heuristics results in a considerable improvement in the quality of sparse matrix orderings over conventional ordering methods. In addition, our graph-partitioning based ordering algorithm is more parallelizable than minimum-degree based orderings algorithms and it renders the ordered matrix more amenable to parallel factorization.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Cleve Ashcraft and Joseph W.-H. Liu. Robust ordering of sparse matrices using multisection. Technical Report CS 96-01, Department of Computer Science, York University, Ontario, Canada, 1996.Google Scholar
  2. 2.
    Cleve Ashcraft and Joseph W.-H. Liu. Generalized nested dissection: Some recent progress. In Proceedings of Fifth SIAM Conference on Applied Linear Algebra, Snowbird, Utah, June 1994.Google Scholar
  3. 3.
    Timothy A. Davis, Patrick Amestoy, and Iain S. Duff. An approximate minimum degree ordering algorithm. Technical Report TR-94-039, Computer and Information Sciences Department, University of Florida, Gainesville, FL, 1994.Google Scholar
  4. 4.
    Iain S. Duff and Torbjorn Wiberg. Remarks on implementations of O(n 1/2 τ) assignment algorithms. ACM Transactions on Mathematical Software, 14:267–287, 1988.Google Scholar
  5. 5.
    C. M. Fiduccia and R. M. Mattheyses. A linear time heuristic for improving network partitions. In Proceedings of the 19th IEEE Design Automation Conference, pages 175–181, 1982.Google Scholar
  6. 6.
    Alan George and Joseph W.-H. Liu. Computer Solution of Large Sparse Positive Definite Systems. Prentice-Hall, Englewood Cliffs, NJ, 1981.Google Scholar
  7. 7.
    Alan George and Joseph W.-H. Liu. The evolution of the minimum degree ordering algorithm. SIAM Review, 31(1):1–19, March 1989.Google Scholar
  8. 8.
    Madhurima Ghose and Edward Rothberg. A parallel implementation of the multiple minimum degree ordering heuristic. Technical report, Old Dominion University, Norfolk, VA, 1994.Google Scholar
  9. 9.
    Anshul Gupta. Fast and effective algorithms for graph partitioning and sparse matrix reordering. Technical Report RC 20496 (90799), IBM T. J. Watson Research Center, Yorktown Heights, NY, July 10, 1996. Available on the WWW at the IBM Research Cyber Journal site at http://www.research.ibm.com:8080/.Google Scholar
  10. 10.
    Anshul Gupta. WGPP: Watson graph partitioning (and sparse matrix ordering) package: Users manual. Technical Report RC 20453 (90427), IBM T. J. Watson Research Center, Yorktown Heights, NY, May 6, 1996.Google Scholar
  11. 11.
    Anshul Gupta. Graph partitioning based sparse matrix ordering algorithms for finite-element and optimization problems. In Proceedings of the Second SIAM Conference on Sparse Matrices, October 1996.Google Scholar
  12. 12.
    Anshul Gupta, George Karypis, and Vipin Kumar. Highly scalable parallel algorithms for sparse matrix factorization. Technical Report 94-63, Department of Computer Science, University of Minnesota, Minneapolis, MN, 1994. To appear in IEEE Transactions on Parallel and Distributed Systems, 1997. Postscript file available via anonymous FTP from the site ftp://ftp.cs.umn.edu/users/kumar.Google Scholar
  13. 13.
    Anshul Gupta and Vipin Kumar. Parallel algorithms for forward and back substitution in direct solution of sparse linear systems. In Supercomputing '95 Proceedings, December 1995.Google Scholar
  14. 14.
    Bruce Hendrickson and Robert Leland. A multilevel algorithm for partitioning graphs. In Supercomputing '95 Proceedings, 1995. Also available a Technical Report SAND93-1301, Sandia National Laboratories, Albuquerque, NM.Google Scholar
  15. 15.
    Bruce Hendrickson and Edward Rothberg. Improving the runtime and quality of nested dissection ordering. Technical Report SAND96-0868J, Sandia National Laboratories, Albuquerque, NM, 1996.Google Scholar
  16. 16.
    George Karypis, Anshul Gupta, and Vipin Kumar. Parallel formulation of interior point algorithms. Technical Report 94-20, Department of Computer Science, University of Minnesota, Minneapolis, MN, April 1994. A short version appears in Supercomputing '94 Proceedings.Google Scholar
  17. 17.
    George Karypis and Vipin Kumar. METIS: Unstructured graph partitioning and sparse matrix ordering system. Technical report, Department of Computer Science, University of Minnesota, 1995.Google Scholar
  18. 18.
    George Karypis and Vipin Kumar. Multilevel k-way partitioning scheme for irregular graphs. Technical Report TR 95-064, Department of Computer Science, University of Minnesota, 1995.Google Scholar
  19. 19.
    George Karypis and Vipin Kumar. Parallel multilevel graph partitioning. Technical Report TR 95-036, Department of Computer Science, University of Minnesota 1995.Google Scholar
  20. 20.
    B. W. Kernighan and S. Lin. An efficient heuristic procedure for partitioning graphs. The Bell System Technical Journal, 1970.Google Scholar
  21. 21.
    Vipin Kumar, Ananth Grama, Anshul Gupta, and George Karypis. Introduction to Parallel Computing: Design and Analysis of Algorithms. Benjamin/Cummings Redwood City, CA, 1994.Google Scholar
  22. 22.
    Alex Pothen and C.-J. Fan. Computing the block triangular form of a sparse matrix. ACM Transactions on Mathematical Software, 1990.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Anshul Gupta
    • 1
  • Fred Gustavson
    • 1
  1. 1.IBM T.J. Watson Research CenterYorktown HeightsUSA

Personalised recommendations