Advertisement

Journal of Computer Science and Technology

, Volume 3, Issue 4, pp 263–272 | Cite as

Optimization for the parallel execution of non-DO loops under Leading Iteration Model

  • Chen Haibo 
  • Ci Yungui 
Regular Papers
  • 15 Downloads

Abstract

In this paper, we present a group of optimization rules to systematically improve the parallelism of non-DO loops under Leading Iteration Model. Control relations in programs are classified into three schemata, which can be conveniently represented with flow charts. Nine optimization rules are presented based on the flow chart description of Program Control Schemata, and they are used to show how to improve the parallelism of non-DO loops with examples. Finally, problems are discussed and further research efforts are described.

Keywords

Parallel Execution Control Relation Generate Synchronization Program Transformation Optimization Rule 
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. [1]
    M.J. Wolfe, Optimizing Supercompilers for Supercomputers, PH. D. Thesis, University of Illinois at Urbana-Champaign, 1982.Google Scholar
  2. [2]
    C.D. Polychronopoulos, D.J. Kuck and D.A. Padua, Execution of Parallel Loops on Parallel Processor Systems, Proc. of the 1986 ICPP.Google Scholar
  3. [3]
    S.P. Midkiff and D.A. Padua, Compiler Generated Synchronization for DO Loops, Proc. of the 1986 ICPP.Google Scholar
  4. [4]
    R. Cytron, Doacross: Beyond Vectorization for Multiprocessors, Proc. of the 1986 ICPP.Google Scholar
  5. [5]
    D. Callahan, A Global Approach to Detection of Parallelism, Rice COMP TR86-40, May 1986.Google Scholar
  6. [6]
    R. Stron and S. Yemini, Synthesizing Distributed and Parallel Programs through Optimistic Transformations, Proc. of the 1985 ICPP.Google Scholar
  7. [7]
    Fan Zhihua, The Tracing Subscript Method for Vectorization of Serial Arithmetic,Scientia Sinica (Series A),26:8(1983).Google Scholar
  8. [8]
    Guo Qiang and Hai-bo Chen, The relevence analysis approach to vectorization,Chinese Journal of Computers,8:5(1985).Google Scholar
  9. [9]
    Guo Qiang and Hai-bo Chen, Interval segmentation techniques in vectorization,Chinese Journal of Computers,8:5(1985).Google Scholar
  10. [10]
    Hai-Bo Chen, Yun-Gui Ci, Parallel Execution of Non-DO loops, Proc. of the 1987 ICPP.Google Scholar
  11. [11]
    Hai-Bo Chen, Yun-Gui Ci, Speedup Studies of Leading Iteraction Model for Non-DO Loops in Comparison with Sequential Iteration Model, CIT Comp TR87-6103, May 1987.Google Scholar

Copyright information

© Science Press, Beijing China and Allerton Press Inc. 1988

Authors and Affiliations

  • Chen Haibo 
    • 1
  • Ci Yungui 
    • 1
  1. 1.Changsha Institute of TechnologyChangshaChina

Personalised recommendations