Skip to main content
SpringerLink
Log in

Coarse-grain parallelizations of interval algorithms decomposing dense matrices and solving triangular systems on multicomputers

Grobkörnige Parallelisierungen von Intervallalgorithmen zur Faktorisierung vollbesetzter Matrizen und zur Lösung von Dreieckssystemen auf Multicomputern

  • Published:
Computing Aims and scope Submit manuscript

Abstract

The evaluation of scalar products with maximum accuracy plays an important role in computing inclusions for the solutions of linear systems. In this paper, we discuss this operation within the context of parallel algorithms for distributed-memory systems (multicomputers). We describe new variants for solving triangular systems of linear equations and for computing the LU factorization of matrices under the assumption that scalar products are implemented as single, indivisible operations and that no processor works on different scalar products simultaneously. All algorithms work in the real and interval case; the theoretical results are supplemented by measurements obtained from a transputer network.

Zusammenfassung

Die Auswertung von Skalarprodukten mit maximaler Genauigkeit spielt eine wichtige Rolle bei der Berechnung von Lösungseinschließungen für lineare Gleichungssysteme. In dieser Arbeit diskutieren wir diese Operation im Zusammenhang mit parallelen Algorithmen für speicherentkoppelte Systeme (Multicomputer). Wir beschreiben neue Varianten zur Lösung linearer Dreieckssysteme und zur Berechnung der LU-Zerlegung unter der Annahme, daß Skalarprodukte als unteilbare Operationen implementiert sind und daß kein Prozessor an mehreren Skalarprodukten gleichzeitig arbeitet. Alle Algorithmen sind für Punkt- und Intervallprobleme anwendbar; die theoretischen Resultate werden durch Messungen auf einem Transputernetzwerk bestätigt.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. Alefeld, G., Herzberger, J.: Introduction to interval computations. New York: Academic Press 1983.

    Google Scholar 

  2. Atanassova, L., Herzberger, J. (eds): Computer arithmetic and enclosure methods. Amsterdam: Elsevier 1992.

    Google Scholar 

  3. Bohlender, G.: What do we need beyond IEEE arithmetic. In: Computer arithmetic and self-validating numerical methods (Ullrich, C. P. ed.), pp. 1–32. New York: Academic Press 1990.

    Google Scholar 

  4. Dunigan, T.: Performance of the INTEL iPSC/860 and NCUBE 6400 hypercubes. Technical Report ORNL/TN-11790, Oak Ridge National Laboratory, 1991.

  5. Eisenstat, S. C., Heath, M. T., Henkel, C. S., Romine, C. H.: Modified cyclic algorithms for solving triangular systems on distributed-memory multiprocessors. SIAM J. Sci. Stat. Comput., 589–600 (1988).

  6. Heath, M., Romine, C.: Parallel solution of triangular systems on distributed-memory multi-processors. SIAM J. Sci. Stat. Comput.9, 558–588 (1988).

    Article  MathSciNet  Google Scholar 

  7. Klein, W.: Verified results for linear systems with sparse matrices. In: Accurate numerical algorithms—A collection of research papers (Ullrich, C. P., Wolff von Gudenberg, J. eds), pp. 137–161. Berlin Heidelberg New York Tokyo: Springer 1989.

    Google Scholar 

  8. Kuck, D.: Parallel processing of ordinary programs. Adv. Comput.15, 119–179 (1976).

    Google Scholar 

  9. Kulisch, U., Miranker, W. (eds.): A new approach to scientific computation. New York: Academic Press 1983.

    Google Scholar 

  10. Ortega, J. M.: Theijk forms of factorization methods I. Vector computers. Parallel Comput.7, 135–147 (1988).

    MATH  MathSciNet  Google Scholar 

  11. Ortegra, J. M.: Theijk forms of factorization methods II. Parallel systems. Parallel Comput.7, 149–162 (1988).

    MathSciNet  Google Scholar 

  12. Quinn, M.: Designing efficient algorithms for parallel computers New York: McGraw-Hill 1987.

    Google Scholar 

  13. Reith, R.: Wissenschaftliches Rechnen auf Multicomputern—BLAS-Routinen und die Lösung linearer Gleichungssysteme mit Fehlerkontrolle. Dissertation, Institut für Informatik, Universität Basel, 1993.

  14. Reith, R.: Description and interfaces of accurate BLAS-routines for transputer networks. Technical Report 93-5, Institut für Informatik, Universität Basel, 1993.

  15. Romine, C.: Factorization methods for the parallel solution of linear systems. Ph. D. Thesis, Applied Mathematics, University of Virginia, 1986.

  16. Ullrich, C. P., Reith, R.: A reliable linear algebra library for transputer networks. Proceedings of the International Congress on Computer Systems and Applied Mathematics, CSAM 93, St. Petersburg, July 19–23 (1993).

Download references

Author information

Authors and Affiliations

  1. Institut für Informatik, Universität Basel, Mittlere Strasse 142, CH-4056, Basel

    R. Reith &  C. P. Ullrich

Authors
  1. R. Reith
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. C. P. Ullrich
    View author publications

    You can also search for this author in PubMed Google Scholar

Rights and permissions

Reprints and permissions

About this article

Cite this article

Reith, R., Ullrich, C.P. Coarse-grain parallelizations of interval algorithms decomposing dense matrices and solving triangular systems on multicomputers. Computing 53, 243–257 (1994). https://doi.org/10.1007/BF02307377

Download citation

  • Received:

  • Revised:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF02307377

AMS Subject Classifications

Key words

Search

Navigation