Parallel game tree search on SIMD machines

  • Holger Hopp
  • Peter Sanders
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 980)


We describe an approach to the parallelization of game tree search on SIMD machines. It turns out that the single-instruction restriction of SIMD-machines is not a big obstacle for achieving efficiency. We achieve speedups up to 5850 on a 16K processor MasPar MP-1 if the search trees are sufficiently large and if there are no strong move ordering heuristics. To our best knowledge, the largest speedups previously reported (usually on MIMD machines) are more than an order of magnitude smaller.


Parallel game tree search load balancing program transformations for SIMD 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    R. D. Blumofe and C. E. Leiserson. Scheduling multithreaded computations by work stealing. In Foundations of Computer Science, 1994.Google Scholar
  2. 2.
    V. Cung and L. Gotte. A first step towards the massively parallel game-tree search. In International Workshop on Parallel Processing for Artificial Intelligence, pages 88–93, Chambery, 1993. Elsevier.Google Scholar
  3. 3.
    O. I. El-Dessouki and W. H. Huen. Distributed enumeration on between computers. IEEE Transactions on Computers, C-29(9):818–825, September 1980.Google Scholar
  4. 4.
    R. Feldmann. Game Tree Search on Massively Parallel Systems. PhD thesis, Universität Paderborn, August 1993.Google Scholar
  5. 5.
    R. Feldmann, P. Mysliwietz, and B. Monien. Studying overheads in massively parallel min/max-tree evaluation. In ACM Symposium on Parallel Architectures and Algorithms, pages 94–103, 1994.Google Scholar
  6. 6.
    R. Finkel and J. Fishburn. Parallelism in alpha-beta search. Artificial Intelligence, 19:89–106, 1982.CrossRefGoogle Scholar
  7. 7.
    W. D. Hillis. The Connection Machine. Series in Artificial Inteligence. MIT Press, Cambridge, MA, 1985.Google Scholar
  8. 8.
    H. Hopp. Parallele Spielbaumsuche auf SIMD-Rechnern. Diplomarbeit, Universität Karlsruhe, Feb. 1995.Google Scholar
  9. 9.
    C. F. Joerg and B. C. Kuszmaul. Massively parallel chess. In Third DIMACS parallel implementation challenge workshop, pages 299–308. Rutgers University, 1994.Google Scholar
  10. 10.
    G. Karypis and V. Kumar. Unstructured tree search on SIMD parallel computers. IEEE Transactions on Parallel and Distributed Systems, 5(10):1057–1072, 1994.CrossRefGoogle Scholar
  11. 11.
    D. Knuth and W. Moore. An analysis of alpha-beta pruning. Artificial Intelligence, 6(4):293–326, 1975.CrossRefGoogle Scholar
  12. 12.
    V. Kumar, A. Grama, A. Gupta, and G. Karypis. Introduction to Parallel Computing. Design and Analysis of Algorithms. Benjamin/Cummings, 1994.Google Scholar
  13. 13.
    T. A. Marsland and M. Campbell. Parallel search of strongly ordered game trees. ACM Computing Surveys, 14(4):533–551, Dec. 1982.CrossRefGoogle Scholar
  14. 14.
    MasPar Corporation. MasPar System Overview, July 1992.Google Scholar
  15. 15.
    MasPar Corporation. MPL Reference Manual, May 1993.Google Scholar
  16. 16.
    C. Powley, C. Ferguson, and R. E. Korf. Depth-fist heuristic search on a SIMD machine. Artificial Intelligence, 60:199–242, 1993.CrossRefGoogle Scholar
  17. 17.
    A. Reinefeld. Spielbaum-Suchverfahren. Informatik-Fachberichte, Band 200. Springer-Verlag, 1989.Google Scholar
  18. 18.
    P. Sanders. A detailed analysis of random polling dynamic load balancing. In International Symposium on Parallel Architectures Algorithms and Networks, pages 382–389, Kanazawa, Japan, 1994. IEEE.Google Scholar
  19. 19.
    P. Sanders. Efficient emulation of MIMD behavior on SIMD machines. Technical Report IB 29/95, Universität Karlsruhe, Fakultät für Informatik, 1995.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1995

Authors and Affiliations

  • Holger Hopp
    • 1
  • Peter Sanders
    • 1
  1. 1.University of KarlsruheKarlsruheGermany

Personalised recommendations