Advertisement

Elastic load-balancing for image processing algorithms

  • Serge Miguet
  • Yves Robert
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 591)

Abstract

In this paper, we introduce a data redistribution algorithm which aims at dynamically balancing the workload of image processing algorithms on distributed memory processors. First we briefly review state-of-the-art techniques for load balancing application-specific algorithms. Then we describe the data redistribution technique, which we term “elastic load balancing” in a general framework. We demonstrate the usefulness of our redistribution strategy by comparing the efficiency obtained with and without the elastic algorithm for a thinning algorithm which aims at extracting the skeleton of a binary image. We report experimental results obtained with a Supernode machine, based upon reconfigurable networks of 32 Transputers [Nic]. We obtain a speedup of up to 28 over the sequential algorithm, using a Mandelbrot set as a test image. Note that the speedup with a static allocation of the picture was limited to 17 with the same test image, due to the load imbalance among the processors.

Key-words

load balancing data partitioning data redistribution dynamic allocation image processing algorithms distributed memory processors 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [BBa]
    Belkhale K.P., Banerjee P., “Recursive partitions on multiprocessors”, in The Fifth Distributed Memory Conference, D.W. Walker et al. eds., IEEE Computer Society Press (1990), 930–938Google Scholar
  2. [BBo]
    Berger M.J., Bokhari S.H., “A partitioning strategy for nonuniform problems on multiprocessors”, IEEE Trans. Computers 36, 5 (1987), 570–580Google Scholar
  3. [Bok]
    Bokhari S.H., “Partitioning problems in parallel, pipelined and distributed computing”, IEEE Trans. Computers 37, 1 (1988), 48–57CrossRefGoogle Scholar
  4. [BT]
    Baek J.H., Teath K.A., “Parallel thinning on a distributed memory machine”, in The Fifth Distributed Memory Conference, D.W. Walker et al. eds., IEEE Computer Society Press (1990), 72–75Google Scholar
  5. [Cyb]
    Cybenko G., “Dynamic load balancing for distributed memory multiprocessors”, J. Parallel Distributed Computing 7 (1989), 279–301CrossRefGoogle Scholar
  6. [Dun]
    Dunigan, T.H., “Hypercube performance”, in Hypercube Multiprocessors 1987, H.T. Heath ed., SIAM Press (1987), 178–192Google Scholar
  7. [EJ]
    Embrechts H., Jones J.P., “An input/output algorithm for M-dimensional decompositions on N-dimensional hypercubes multicomputers”, in The Fifth Distributed Memory Conference, D.W. Walker et al. eds., IEEE Computer Society Press (1990), 876–882Google Scholar
  8. [GH]
    Guo Z., Hall R.W., “Parallel thinning with two subiteration algorithms”, C.A.C.M. 32, 3 (1989), 359–373Google Scholar
  9. [Hin]
    Hinz D.Y., “A run-time load balancing strategy for highly parallel systems”, in The Fifth Distributed Memory Conference, D.W. Walker et al. eds., IEEE Computer Society Press (1990), 951–961Google Scholar
  10. [LMR]
    J. J. Li, S.Miguet, Y. Robert, S. Ubeda. “Image processing algorithms on distributed memory machines”, in Parallelism in image Processing. J.C. Simon ed. North Holland, to appear.Google Scholar
  11. [MV]
    MacBryan O.A., Van de Velde E.F., “Hypercube algorithms and implementations”, SIAM J. Sci. Stat. Comput. 8, 2 (1987), s227–s287CrossRefGoogle Scholar
  12. [Nic]
    Nicole D. A., “Esprit Project 1085, Reconfigurable Transputer Processor Architecture”, in CONPAR 88, C. R. Jesshope et al. eds., Cambridge University Press (1989), 81–89.Google Scholar
  13. [Ols]
    Olszewski J., “A flexible thinning algorithm allowing parallel, sequential and distributed application,” Technical report, Unviversität des Saarlandes, Saarbrücken, Germany (1990), to appear in ACM Trans. Math. SoftwareGoogle Scholar
  14. [Rob]
    Robert Y, The impact of vector and parallel architectures on the Gaussian elimination algorithm, Manchester University Press and John Wiley (1990)Google Scholar
  15. [RW]
    Rosing M., Weaver R.P., “Mapping data to processors in distributed memory computations”, in The Fifth Distributed Memory Conference, D.W. Walker et al. eds., IEEE Computer Society Press (1990), 884–893Google Scholar
  16. [SE]
    Sadayappan P., Ercal F., “Nearest-neighbor mappings of finite element graphs onto processor meshes”, IEEE Trans. Computers 36, 12(1987), 1408–1424Google Scholar
  17. [SSc1]
    Saad, Y., & Schultz, M.H., “Topological properties of hypercubes”, IEEE Trans. Computers 37, 7 (1988), 867–872CrossRefGoogle Scholar
  18. [SSc2]
    Saad, Y. & Schultz, M.H., “Data communication in parallel architectures”, Parallel Computing 11 (1989), 131–150MathSciNetGoogle Scholar
  19. [SSo]
    Snyder L., Socha D.G., “An algorithm producing balanced partitionings of data arrays”, in The Fifth Distributed Memory Conference, D.W. Walker et al. eds., IEEE Computer Society Press (1990), 867–875Google Scholar
  20. [Ube]
    Ubeda S., “Comparison of thinning algorithm on distributed memory machines”. Technical report LIP-IMAG 90–29 (1990). Submitted for publicationGoogle Scholar
  21. [XH]
    Xu J., Hwang K., “Heuristic methods for dynamic load balancing in a messagepassing supercomputer”, in Supercomputing'90, IEEE Computer Society Press (1990), 888–897Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1992

Authors and Affiliations

  • Serge Miguet
    • 1
  • Yves Robert
    • 1
  1. 1.Laboratoire de l'Informatique du Parallélisme LIP-IMAGEcole Normale Supérieure de LyonLyon Cedex 07France

Personalised recommendations