Refined fixed granularity algorithm on Networks of Workstations

NOW 환경에서 개선된 고정 분할 단위 알고리즘

  • Gu, Bon-Geun (Dept. of Computer Engineering, Chungju National University)
  • 구본근 (충주대학교 컴퓨터공학과)
  • Published : 2001.06.01

Abstract

At NOW (Networks Of Workstations), the load sharing is very important role for improving the performance. The known load sharing strategy is fixed-granularity, variable-granularity and adaptive-granularity. The variable-granularity algorithm is sensitive to the various parameters. But Send algorithm, which implements the fixed-granularity strategy, is robust to task granularity. And the performance difference between Send and variable-granularity algorithm is not substantial. But, in Send algorithm, the computing time and the communication time are not overlapped. Therefore, long latency time at the network has influence on the execution time of the parallel program. In this paper, we propose the preSend algorithm. In the preSend algorithm, the master node can send the data to the slave nodes in advance without the waiting for partial results from the slaves. As the master node sent the next data to the slaves in advance, the slave nodes can process the data without the idle time. As stated above, the preSend algorithm can overlap the computing time and the communication time. Therefore we reduce the influence of the long latency time at the network and the execution time of the parallel program on the NOW. To compare the execution time of two algorithms, we use the $320{\times}320$ matrix multiplication. The comparison results of execution times show that the preSend algorithm has the shorter execution time than the Send algorithm.

Keywords

References

  1. B. Wilkinson and M. Allen, 'PARALLEL PROGRAMMING : Technique and Applications Using Networked Workstation and Parallel Computers,' Prentice Hall, 1999
  2. A Piotrowski and S. Dandamudi, 'A Comparative Study of Load Sharing on Networks of Workstations,' Proc. Int. Conf. Parallel and Distributed Computing System, New Orleans, Oct. 1997
  3. T. Anderson, D. Culler, D. Patterson and the NOW team, 'A Case for NOW(Networks of Workstations),' IEEE Micro, 15(2), pp.54-64, Feb. 1995 https://doi.org/10.1109/40.342018
  4. A. Piotrowski and S. Dandamudi, 'Performance of a Parallel Application on Network of Workstations,' 11th Int. Symp. High Performance Computing Systems, Winnipeg, pp.429-440, July. 1997
  5. C. Polychronopoulos and D. Kuck, 'Guided Self Scheduling : A Practical Scheduling Scheme for Parallel Computers,' IEEE Trans. Computers, Vol.C-36, No.12, pp.1425-1439, Dec. 1987
  6. S. Hummel, E. Schonberg and L. Flynn, 'Factoring : A Method for Scheduling Parallel Loops,' Comm. ACM, Vol.35, No.8, pp.90-101, Aug. 1992 https://doi.org/10.1145/135226.135232
  7. T. Tzen and L. Ni, 'Dynamic Loop Scheduling for Shared Memory Multiprocessors,' IEEE Trans. Parallel Dist. Syst., Vol.4, No.1, pp.87-98, Jan. 1993 https://doi.org/10.1109/71.205655
  8. J. Jacob and S. Lee, 'Task Spreading and Shrinking on Multiprocessor Systems and Networks of Workstations,' IEEE Trans. Parallel and Distributed systems, Vol.10, No.10, pp.1082-1101, Oct. 1999 https://doi.org/10.1109/71.808157