High Performance Data Parallel Recursion

  • Norbert Duppel
Conference paper


The purpose of this paper is twofold. First we give results of a high performance implementation of data parallel recursion. Two algorithms demonstrate a linear growth in the range from 1 to 15 CPUs, and there is reason to believe that this will continue up to 45 resp. 25 CPUs.

The first algorithm is usable for arbitrarily large amounts of data, for the faster second one the result has to fit in the main memory available on all CPUs together. These algorithms can be perfectly used for transitive closure processing, and also for generalized transitive closures as e.g. billof-materials.

Second we give an algorithm for the integration of this data parallel algorithms in a general Deductive Database. We demonstrate the compilation of arbitrary recursive Horn logic rules into our data parallel algorithm. In fact we have just finished the integration in the Parallel Deductive Database developed at Stuttgart university.


Transitive Closure Server Class Main Program Rule System Database Operation 
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]
    Dayal, U., Smith, J.; PROBE: A Knowledge-Oriented DBMS, Proc. Islamorada Workshop On Large Scale Knowledgebase And Reasoning Systems, 1985Google Scholar
  2. [2]
    Ullman, J. D.; Implementation of Logical Query Languages for Databases, ACM Transactions on Database Systems, Vol. 10, No. 3, pp 298–321, 1985CrossRefGoogle Scholar
  3. [3]
    Lefebvre, A., Vieille, L.; On Deductive Query Evaluation in the DedGin System, Proc. 1st Int. Conf. on Deductive and O-O Databases, 1989Google Scholar
  4. [4]
    Boral, H.; Parallelism in BUBBA, Proc. Int. Symposium on Databases in Parallel and Distributed Systems, Austin, Tx, 1988Google Scholar
  5. [5]
    Duppel, N., Gugel, D., Maier, J., Reuter, A., Schiele, G.; Proc. Workshop on PROSPECT, University of Stuttgart, 1990Google Scholar
  6. [6]
    Bayer, R.; Query Evaluation and Recursion in Deductive Database Systems, Technical University of Munich, internal review, 1985Google Scholar
  7. [7]
    Bancilhon, F. and Ramakrishnan, R.; An Amateur’s Introduction to Recursive Query Processing Strategies, ACM SIG-MOD, Vol. 15, No. 2, p. 16, 1986CrossRefGoogle Scholar
  8. [8]
    Bancilhon, F., Maier, D., Sagiv, Y., Ullman, J.; Magic sets and other strange ways to implement logic programs, Proc. 5th ACM SIGMOD-SIGACT PODS, 1986Google Scholar
  9. [9]
    Beeri, C., Ramakrishnan, R.; On the power of Magic, MCC Technical Report DB-061–86, 1986Google Scholar
  10. [10]
    Sacca, D., Zaniolo, C.; Magic Counting Methods, Proc. ACM SIGMOD, 1987Google Scholar
  11. [11]
    Gallaire, H., Minker, M.; Logic and Databases: A Deductive Approach, ACM Computing Surveys, Vol. 16, No. 2, 1984Google Scholar
  12. [12]
    Schiele, G.; Eine Testumgebung zur Untersuchung paralleler Verarbeitungsstrategien in komplexen Transaktionen, Informatik Fachberichte, No. 204, Springer Verlag, 1989Google Scholar
  13. [13]
    Gugel, Dieter; Verteilte parallele Verarbeitung komplexer Transaktionen - Entwurf geeigneter Basismechanismen, Diplomarbeit am Institut für Informatik, Universität Stuttgart, 1989Google Scholar
  14. [14]
    Duppel, N.; Implementation of a Parallel Deductive Database, Proc. GFKI 1991, Salzburg, 1991Google Scholar
  15. [15]
    Staiger, R.; Massively parallel bill-of-material processing (in German), Studienarbeit No. 844, IPVR, university of Stuttgart, 1990Google Scholar

Copyright information

© Springer-Verlag Wien 1991

Authors and Affiliations

  • Norbert Duppel
    • 1
  1. 1.University of Stuttgart, IPVRStuttgartGermany

Personalised recommendations