Advertisement

A Parallel Implementation of the Eigenproblem for Large, Symmetric and Sparse Matrices

  • E. M. Garzón
  • I. García
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1697)

Abstract

This work studies the eigenproblem of large, sparse and symmetric matrices through algorithms implemented in distributed memory multiprocessor architectures. The implemented parallel algorithm operates in three stages: structuring input matrix (Lanczos Method), computing eigenvalues (Sturm Sequence) and computing eigenvectors (Inverse Iteration). Parallel implementation has been carried out using a SPMD programming model and the PVM standard library. Algorithms have been tested in a multiprocessor system Cray T3E. Speed-up, load balance, cache faults and profile are discussed. From this study, it follows that for large input matrices our parallel implementations perceptibly improve the management of the memory hierarchy.

Keywords

Parallel Algorithm Parallel Implementation Input Matrix Test Matrice Memory Hierarchy 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Basermann, A.; Weidner, P. A parallel algorithm for determining all eigenvalues of large real symmetric tridiagonal matrices. Parallel Computing 18, pag 1129–1141. 1992.zbMATHCrossRefMathSciNetGoogle Scholar
  2. 2.
    Bernstein, H.J. An Accelerated Bisection Method for the Calculation of Eigenvalues of a Symmetric Tridiagonal Matrix. Numer. Math. 43, pag. 153–160. 1984.zbMATHCrossRefMathSciNetGoogle Scholar
  3. 3.
    Berry, M.W. Large-Scale Sparse Singular Value Computations. The International Journal of Supercomputer. Vol 6, pg. 171–186. 1992.Google Scholar
  4. 4.
    Cullum, J.K. and Willoughby, R.A. Lanczos Algorithms for Large Symmetric Eigenvalue Computations. Vol. I Theory. Birkhauser Boston, Inc., 1985.Google Scholar
  5. 5.
    Duff, I.S.; Grimes, R.G. and Lewis, J.G. User’s Guide for the Harwell-Boeing Sparse Matrix Collection. Research and Technology Division, Boeing Computer Services. 1992.Google Scholar
  6. 6.
    García, I; Garzón, E.M.; Cabaleiro, J.C.; Carazo, J.M. and Zapata, E.L. Parallel Tridiagonalization of Symmetric Matrices Based on Lanczos Method. Parallel Computing and Transputer Applications. Vol I, 236–245, 1992.Google Scholar
  7. 7.
    Garzón, E.M. and García, I. Parallel Implementation of the Lanczos Method for Sparse Matrices: Analysis of the Data Distributions. Proceedings ICS’96. p. 294–300. Philadelphia, 1996.Google Scholar
  8. 8.
    Garzón, E.M. and García, I. Evaluation of the Work Load Balance in Irregular Problems Using Value Based Data Distributions. Proceedings of Euro-PDS’97. p. 137–143. 1997.Google Scholar
  9. 9.
    Golub, G.H. and Van Loan C. F. Matrix Computation. The Johns Hopkins University Press., Baltimore and London., 1993.Google Scholar
  10. 10.
    Simon, H.D. Analysis of the symmetric Lanczos Algorithm with Reorthogonalization Methods. Linear Algebra and its Applications. Vol 61, pg. 101–131. 1984.zbMATHCrossRefMathSciNetGoogle Scholar
  11. 11.
    Sy-Shin Lo, Bernard Philippe and Ahmed Samed. A multiprocessor Algorithm for the Symmetric Tridiagonal Eigenvalue Problem. SIAM J. Sci. Stat. Comput. vol-8, No. 2, March 1987.Google Scholar
  12. 12.
    Wilkinson, J.H. The Algebraic Eigenvalue Problem. Clarendon Press. Oxford, 1965.zbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • E. M. Garzón
    • 1
  • I. García
    • 1
  1. 1.Dpto Arquitectura de Computadores y ElectrónicaUniversidad de AlmeríaAlmeríaSpain

Personalised recommendations