Advertisement

The gene concept and its implementation for a dataflow schemed parallel computer

  • Kenji Toda
  • Yoshinobu Uchibori
  • Toshitsugu Yuba
Submitted Presentations
Part of the Lecture Notes in Computer Science book series (LNCS, volume 365)

Abstract

A dataflow scheme is suitable for multi-processor systems to extract parallelism naturally, but mapping ideal parallel computations to limited execution resources is a major problem. The Gene concept is proposed to provide flexible control of parallelism in dataflow schemed parallel computers. A gene is the property carried by data and propagated from ancestor operations to descendant operations. In this way, the Gene groups the operations according to properties. By checking the properties, the Gene can cease and suspend the execution of operation groups, and control the priority of execution among groups. These functions are essential for general purpose highly parallel computers allowing multi-programming, multi-user and standalone usage. This paper proposes the Gene concept and discusses its implementation and usage, then shows its effectiveness by simulations.

Keywords

Execution Time Function Call Priority Queue Priority Structure Speedup Curve 
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. [Arv80]
    Arvind, Kathail V., and Pingali K., "A Data Flow Architecture with Tagged Tokens", TM-174, Lab. Comp. Sci., MIT, 1980.Google Scholar
  2. [Arv86]
    Arvind, and Culler, David E., "Managing Resources in a Parallel Machine", Proc. of IFIP TC10 WC, Fifth Generation Computer Architectures, J.V.Woods(ed.), Elsevier Science Publishers B.V.(North-Holland), pp.103–121, 1986Google Scholar
  3. [Kuc86]
    Kuck, David J., Davidson, Edward S., Lawrie, Duncan H., and Sameh Ahmed H., "Parallel Supercomputing Today and the Cedar Approach", Science 231, pp.967–974, 1986.Google Scholar
  4. [Bab84]
    Babb R. G. II, "Parallel Processing with large-Grain Data Flow Techniques", Computer, Vol.17, No.7, pp.55–61, 1984.Google Scholar
  5. [Tod87]
    Toda, Kenji, Uchibori, Yoshinobu, and Yuba, Toshitsugu, "A Macro-Dataflow Model and Its Language", Preprint of WGARC IPS Japan, 66–6, 1987. (in Japanese)Google Scholar
  6. [Rug87]
    Ruggiero, Carlos A., "Throttle Mechanisms for the Manchester Dataflow Machine", Dep. of Comp. Science, Univ. of Manchester, Technical Report UMCS-87-6-1, 1987.Google Scholar
  7. [Bar87]
    Barahona, Pedro M.C.C., "Specification and Control of Execution of Nondeterministic Dataflow Programs", Dep. of Comp. Science, Univ. of Manchester, Technical Report UMCS-87-6-1, 1987.Google Scholar
  8. [Tak87]
    Takesue, Masaru, "A Load Control Mechanism for Data Flow Machines", Tran. of IEICE, Vol. J70-D, No.10, pp.1878–1889, 1987. (in Japanese)Google Scholar
  9. [EM86]
    Toda, Kenji, Yamaguchi, Yoshinori, Uchibori, Yoshinobu, and Yuba, Toshitsugu, "Preliminary Measurements of the ETL LISP-based Data-driven Machine", Proc. of IFIP TC10 WC, Fifth Generation Computer Architectures, J.V.Woods(ed.), Elsevier Science Publishers B.V.(North-Holland), pp.234–253, 1986.Google Scholar
  10. [U3.2b]
    UNIX 3.2 bsd Reference Manual, AT&T Bell Laboratories, 1986.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1989

Authors and Affiliations

  • Kenji Toda
    • 1
  • Yoshinobu Uchibori
    • 1
  • Toshitsugu Yuba
    • 1
  1. 1.Electrotechnical LaboratoryTsukuba, IbarakiJapan

Personalised recommendations