Advertisement

Preconditioning of Blockstructured Linear Systems with Block ILU on Parallel Computers

  • Volker Mehrmann
Part of the Notes on Numerical Fluid Mechanics (NNFM) book series (NNFM, volume 29)

Abstract

We discuss block incomplete factorizations for large sparse block tridiagonal matrices arising for example in the numerical solution of partial differential equations. We describe a procedure to partition the matrices into subsystems that can be easily, efficiently and in a numerically stable way solved on a parallel computer. The partitioning is based on low rank modifications which generalize the approaches used in Divide and Conquer methods. The effect of these preconditioners are demonstrated for two model problems.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    A. BERMAN AND R. PLEMMONS, Nonnegative Matrices in the Mathematical Sciences ,Academic Press, New York, 1979.zbMATHGoogle Scholar
  2. [2]
    S. Bondeli, Divide and Conquer: Parallele Algorithmen zur Loesung tridiagonaler Gle~ ichungssysteme, Dissertation, ETH Zuerich ,Verlag der Fachvereine, Zuerich, 1991.Google Scholar
  3. [3]
    E. Dick an. J. Linden, A multigrid flux-difference splitting method for steady incom pressible Navier-Stokes equations ,Proceedings of the GAMM Conference on Numerical Methods in Fluid Mechanics, Delft, (1989).Google Scholar
  4. [4]
    L. Elsner an. V. Mehrmann, Convergence of block iterative methods for linear systems arising in the numerical solution of Euler equations ,Numer. Math., 59 (1991), pp. 541–560.CrossRefzbMATHMathSciNetGoogle Scholar
  5. [5]
    R. Freund, G. H. Golub, an. N. Nachtigal Iterative solution of linear systems ,Tech. Report na-91–05, Computer Science Dept., Stanford University, Stanford, Ca. 94305, 1991.Google Scholar
  6. [6]
    W. HACKBUSCH, Iterative Lösung großer schwachbesetzter Gleichungssysteme ,Teubner Verlag, Stuttgart, 1991.CrossRefzbMATHGoogle Scholar
  7. [7]
    A. KRECHEL, H. Plum, an. K. Stueben, Solving tridiagonal linear systems in parallel on local memory mimd machines ,tech. report, Gesellschaft f. Mathematik und Datenverarbeitung, St. Augustin (FRG), 1989.Google Scholar
  8. [8]
    V. Mehrmann, Divide and conquer methods for block tridiagonal systems ,tech. report, Institut für Geometrie und Praktische Mathematik, RWTH Aachen, Aachen FRG, 1991.Google Scholar
  9. [9]
    V. Mehrmann, Divide and conquer methods for tridiagonal linear systems ,Notes on Numerical Fluid Mechanics (Parallel Algorithms for Partial Differential Equations), (Proceedings of the Sixth GAMM-Seminar, Kiel, January 19–21, 1990 Wolfgang Hackbusch, Editor), (1991).Google Scholar
  10. [10]
    R. Varga, Matrix Iterative Analysis ,Prentice-Hall, Englewood Cliffs, N.J., 1962.Google Scholar

Copyright information

© Friedr. Vieweg & Sohn Verlagsgesellschaft mbH, Braunschweig/Wiesbaden 1993

Authors and Affiliations

  • Volker Mehrmann
    • 1
  1. 1.Fakultät für MathematikUniversität BielefeldBielefeld 1Deutschland

Personalised recommendations