Advertisement

A distributed shared memory multiprocessor kit with scalable local complexity

  • F. Höpfl
  • J. Schirrmacher
  • M. Trent
Reconfigurable And Scalable Systems
Part of the Lecture Notes in Computer Science book series (LNCS, volume 457)

Abstract

This paper describes a study for a new multiprocessor kit. The general structure is like the Erlangen multiprocessor system DIRMU 25, but the amount of neighbours is not restricted to a fixed number. In message passing systems this concept is realised by clustering. This paper shows, how the cluster-concept can be applied for shared memory machines. After some general aspects of clustering and configuring systems, the structure of a single cluster is explained in detail from a logical point of view. Already some measurements of an experimental implementation of the system have been done and are presented.

Keywords

memory coupling variable restricted neighbourhood clustering 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Händler, W.; Maehle, E.; Wirl, K.: DIRMU Multiprocessor Configurations, Proc. '85 International Conference Parallel Processing, St. Charls, 1985.Google Scholar
  2. [2]
    Maehle, E.: Fault-Tolerant DIRMU Multiprocessor Configurations, Computer Architecture Tech. Com. Newsletter, IEEE Computer Society, June 1985Google Scholar
  3. [3]
    Trent, M.; Schirrmacher, J.: Ein speichergekoppeltes Multiprozessorsystem basierend auf dem Motorola MC 68030, Teil 1 & 2, Diplomarbeiten am IMMD 3 der Universität Erlangen-Nürnberg, 1989Google Scholar
  4. [4]
    Höpfl, F.; Trent, M.; Schirrmacher, J.: Eine objektorientierte Architektur für paralleles LISP, GI-Fachtagung "Architektur von Rechnersystemen", 1990Google Scholar
  5. [5]
    INMOS: "Transputer reference manual", Prentice Hall, 1988Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1990

Authors and Affiliations

  • F. Höpfl
    • 1
  • J. Schirrmacher
    • 1
  • M. Trent
    • 1
  1. 1.Dept. of Computer Science (IMMD III)University of Erlangen-NürnbergErlangenFRG

Personalised recommendations