Message Encoding Techniques for Efficient Arrary Redistribution
Abstract
In this paper, we present message encoding techniques to improve the performance of BLOCK-CYCLIC(kr) to BLOCK-CYCLIC(r) (and vice versa) array redistribution algorithms. The message encoding techniques are machine independent and could be used with different algorithms. By incorporating the techniques in array redistribution algorithms, one can reduce the computation overheads and improve the overall performance of array redistribution algorithms. To evaluate the performance of the techniques, we have implemented the message encoding techniques into some array redistribution algorithms on an IBM SP2 parallel machine. The experimental results show that the execution time of array redistribution algorithms with the message encoding techniques is 3% to 22% faster than those without the message encoding techniques.
References
[1]
S. Chatterjee, J. R. Gilbert, F. J. E. Long, R. Schreiber, and S.-H. Teng, "Generating Local Address and Communication Sets for Data Parallel Programs," JPDC, Vol. 26, pp. 72-84, 1995.
[2]
S. K. S. Gupta, S. D. Kaushik, C.-H. Huang, and P. Sadayappan, "On Compiling Array Expressions for EffIcient Execution on Distributed-Memory Machines," JPDC, Vol. 32, pp. 155-172, 1996.
[3]
S Hiranandani, K. Kennedy, J. Mellor-Crammey, and A. Sethi," Compilation technique for block-cyclic distribution," In Proc. ACM Intl. Conf. on Supercomputing, pp. 392-403, July 1994.
[4]
E. T. Kalns, and L. M. Ni, "Processor Mapping Technique Toward Efficient Data Redistribution, " IEEE TPDS, vol. 6, no. 12, December 1995.
[5]
S. D. Kaushik, C. H. Huang, J. Ramanujam, and P. Sadayappan, "Multiphase array redistribution: Modeling and evaluation," In Proc. of IPPS, pp. 441- 445, 1995.
[6]
K. Kennedy, N. Nedeljkovic, and A. Sethi, "Efficient address generation for block-cyclic distribution," In Proc. of Intl. Conf. on Supercomputing, Barcelona, pp. 180-184, July 1995.
[7]
Y.-W. Lim, Prashanth B. Bhat, and Viktor, K. Prasanna, "Efficient Algorithms for Block-Cyclic Redistribution of Arrays," Proceedings of the Eighth IEEE Symposium on Parallel and Distributed Processing, pp. 74-83, 1996.
[8]
S. Ramaswamy, B. Simons, and P. Banerjee, "Optimization for efficient array Redistribution on Distributed Memory Multicomputers," JPDC, Vol. 38, pp. 217-228, 1996.
[9]
J. M. Stichnoth, D. O'Hallaron, and T. R. Gross," Generating communication for array statements: design, implementation, and evaluation," JPDC, Vol. 21, pp. 150-159, 1994.
[10]
R. Thakur, A. Choudhary, and G. Fox, "Runtime array redistribution in HPF programs, " Proc. 1994 Scalable High Performance Computing Conf., pp. 309-316, May 1994.
[11]
Rajeev. Thakur, Alok. Choudhary, and J. Ramanujam, "Efficient Algorithms for Array Redistribution," IEEE TPDS, vol. 7, no. 6, JUNE 1996.
[12]
A. Thirumalai and J. Ramanujam, "HPF array statements: Communication generation and optimization," 3th workshop on Languages, Compilers and Run-time system for Scalable Computers, Troy, NY, May 1995.
[13]
David W. Walker, Steve W. Otto, "Redistribution of BLOCK-CYCLIC Data Distributions Using MPI," Technical Report ORN/TM-12999, Computer Science and Mathematics Division, Oak Ridge National Laboratory, 1995.
[14]
A. Wakatani and M. Wolfe, "A New Approach to Array Redistribution: Strip Mining Redistribution," In Proc. of Parallel Architectures and Languages Europe, July 1994.
[15]
A. Wakatani and M. Wolfe, "Optimization of array redistribution for distributed memory multicomputer, " In Parallel Computing(submitted), 1994.
Recommendations
Efficient Method for kr->r and r->kr Arrary Redistribution
COMPSAC '97: Proceedings of the 21st International Computer Software and Applications ConferenceArray redistribution is usually required to enhance algorithm performance in many parallel programs on distributed memory multicomputers. Since it is performed at run-time, there is a performance tradeoff between the efficiency of new data decomposition ...
Improving communication scheduling for array redistribution
Many scientific applications require array redistribution when the programs run on distributed memory parallel computers. It is essential to use efficient algorithms for redistribution, otherwise the performance of the programs will degrade ...
Efficient Methods for kr → r and r → kr Array Redistribution1
Array redistribution is usually required to enhance algorithm performance in many parallel programs on distributed memory multicomputers. Since it is performed at run-time, there is a performance tradeoff between the efficiency of new data decomposition for ...
Comments
Information & Contributors
Information
Published In
August 1997
334 pages
ISBN:081868108X
Copyright © Copyright (c) 1997 Institute of Electrical and Electronics Engineers, Inc. All rights reserved.
Publisher
IEEE Computer Society
United States
Publication History
Published: 11 August 1997
Author Tags
Qualifiers
- Article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 0Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0
Reflects downloads up to 29 Jul 2024