Solving Partial Differential Equations on a Network of Transputers

  • E. Verhulst


This paper describes in a first part some numerical experiments using the transputer. A simple equation (two-dimensional Laplace) was chosen as the vehicle for our efforts. This equation was solved using parallelized versions of the classical algorithms: Jacobi, SOR. Two different parallel methods were used for each of the two algorithms, allowing us to draw some more general conclusions regarding such performance-determining parameters as the computation/communication ratio in transputer networks.

We also speculate on the influence of a much faster floating-point unit on the performance of a network for this kind of problems (such a unit is now available in the form of the T800 Transputer).

In a second part an overview is given of an ESPRIT proposal that aims to design and implement a high level user-friendly programming environment for solving PDE on a network of transputers.


Internal Memory Hardware Configuration Package Method Runtime Error Interprocessor Communication 
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.
    . R.W. Hockney and C.R. Jesshope, “Parallel Computers”, Adam Hilger (1981).Google Scholar
  2. The Transputer Reference Manual, INMOS Documentation.Google Scholar
  3. 3.
    . Occam 2 Reference Manual, INMOS Ltd., Prentice Hall (1988).Google Scholar
  4. 4.
    . M. Homewood, The IMS T800 Transputer,IEEE Micro, October 1987.Google Scholar
  5. 5.
    . D. Reed and R. Fujimoto, Multicomputer networks. Message based parallel processing, M.I.T. (1987).Google Scholar

Copyright information

© Plenum Press, New York 1989

Authors and Affiliations

  • E. Verhulst
    • 1
  1. 1.Intelligent Systems InternationalKessel-LOBelgium

Personalised recommendations