Browse > Article
http://dx.doi.org/10.3745/KIPSTA.2002.9A.2.197

Overlapping Effects of Circular Shift Communication and Computation  

Kim, Jung-Hwan (School of Computer Aided Science, Kunkuk University)
Rho, Jung-Kyu (Seokyeong University)
Song, Ha-Yoon (Hongik University)
Publication Information
The KIPS Transactions:PartA / v.9A, no.2, 2002 , pp. 197-206 More about this Journal
Abstract
Many researchers have been interested in the optimization of parallel programs through the latency hiding by overlapping the communication with the computation. We ana1yzed overlapping effects in the circular shift communication which is one of the collective communications being frequently used In many data parallel programs. We measured the time which can be possibly overlapped and the time which cannot be overlapped in over all circular shift communication period on an Ethernet switch-based clustered system. The result from each platform nay be used for the input of optimizing compilers. The previous performance models usually have two kinds of drawbacks one is only based on point-to-point communication, so it is not appropriate for analyzing the overall effects of collective communications. The other provides the performance of collective communication, but no overlapping effect. In this paper we extended the previous models and analyzed the experimental results of the extended model.
Keywords
collective communication; performance model; circular shift; overlapping; data Parallel; cluster;
Citations & Related Records
연도 인용수 순위
  • Reference
1 ZhiWei Xu and Kai Hwang, 'Modeling Communication Overhead : MPI and MPL Performance on the IBM SP2,' IEEE Parallel and Distributed Technology, Vol.4, No.1, pp.9-23, Spring, 1996   DOI   ScienceOn
2 T.Agerwala et al., 'SP2 System Architecture,' IBM Systems Journal, Vol.34, No.2, 1995
3 D. E. Culler, A. Arpaci-Dusseau, R, Arpaci-Dusseau, B. Chun, S. Lumetta, A. Mainwaring, R. Martin, C. Yoshikawa and F. Wong, 'Parallel Computing on the Berkeley NOW,' In Proc. of 9th Joing Symp. on Parallel Processing, Kobe, Japan, 1997
4 T. Sterling et al., 'BEOWULF : A Parallel Worskstation for Scientific Computation,' In Proc. of Int'l Conf. on Parallel Processing, 1995
5 A. Rogers and K. Pingali, 'Process Decomposition though Locality of Reference,' In Proc. of the SIGPLAN '89 Conf. on Programming Language Design and Implementation, June, 1989   DOI
6 C. Koelbel and P. Mehrota, 'Programming Data Parallel Algorithms on Distributed Memory Machine Using Kali,' In Proc. of the 1991 ACM Int'l Conf. on Supercomputing, June, 1991   DOI
7 R.W. Hockney, 'The Communication Challenge for MPP : Intel Paragon and Meiko CS-2,' Parallel Computing, Vol.20, No.3, pp.389-398, March, 1994   DOI   ScienceOn
8 J. M. Squyres et al., 'LAM Installation Guide.'
9 S. Fortune and J. Wyllie, 'Parallelism on Random Access Machines,' In Proc. of the 10th Annual Symp. on Theory of Computing, pp.114-118, 1978   DOI
10 A. Alexandrov et al., 'LogGP : Incorporating Long Messages into the LogP Model,' In Proc. of the 7th Annual Symp. Parallel Algorithms and Architectures, ACM Press, pp.95-105, 1995   DOI
11 D. F. Bacon, S. L. Graham, and O. J. Sharp, 'Compiler Transformations for High-Performance Computing,' ACM Computing Survey, Vol.26, No.4, pp.345-420, December, 1994   DOI   ScienceOn
12 'IBM parallel Environment for AIX : MPL Programming and Subroutine Reference,' GC23-3893-00, IBM Corp.
13 D. E. Culler et al., 'LogP : Towards a Realistic Model of Parallel Computation,' In Proc. of ACM Symp. on Principle and Practice of Parallel Programming, ACM Press, pp.1-12, 1993   DOI
14 MPI Forum, 'MPI : A Message Passing Interface Standard,' Tech. Report CS-94-230, Computer Science Dept.,University of Tennessee, April, 1994
15 A. Geist et al., 'PVM 3 User's Guide and Reference Manual,' September, 1994