Advertisement

Cycle shrinking by dependence reduction

  • Kunio Okuda
Workshop 03 Automatic Parallelization and High Performance Compilers
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1123)

Abstract

We present a new simple cycle shrinking technique called dependence reduction for regular uniform nest algorithm. It consists of a transformation of the dependence graph to reduce the number of execution steps as well as the communication between processors. We gave simple and sufficient conditions for the dependence reduction method to be better than other methods.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Darte, A. and Robert, Y. Mapping uniform loop nests onto distributed memory architectures. Parallel Computing 20(1994) 679–710.CrossRefGoogle Scholar
  2. 2.
    Liu, L. S., Ho, C. W. and Sheu, J. P. On the parallelism of nested for-loops using index shift method. Proc. Internat. Conf. on Parallel Processing (Aug. 1990) II-119–II-123.Google Scholar
  3. 3.
    Okuda, K., Cycle shrinking by dependence reduction. Technical Report, Universidade de São Paulo, 1995.Google Scholar
  4. 4.
    Peir, J. K. and Cytron, R., Minimum distance: a method for partitioning recurrence for multiprocessors. IEEE Trans. Comput. 38(8) (Aug. 1989) 1203–1211.CrossRefGoogle Scholar
  5. 5.
    Polychronopoulos, C. D., Compiler optimization for enhancing parallelism and their impact on architecture design. IEEE Trans. Comput. 37(8) (Aug. 1988) 991–1004.CrossRefGoogle Scholar
  6. 6.
    Robert, Y. and Song, S.W., Revisiting cycle shrinking. Parallel Computing, 18(1992) 481–496.CrossRefGoogle Scholar
  7. 7.
    Shang, W., O'Keefe, M. T. and Fortes, J. A. B. Generalized cycle shrinking. Parallel Algorithms and VLSI Architecture II. P. Quinten and Y. Robert (editors) North Holland, 1991.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1996

Authors and Affiliations

  • Kunio Okuda
    • 1
  1. 1.Universidade de São PauloBrasil

Personalised recommendations