Fumihiko Ino
Abstract
We present a new parallel computational model, named LogGPS, which captures synchronization.
The LogGPS model is an extension of the LogGP model, which abstracts communication on parallel platforms. Although the LogGP model captures long messages with one bandwidth parameter (G), it does not capture synchronization that is needed before sending a long message by high-level communication libraries. Our model has one additional parameter, S, defined as the threshold for message length, above which synchronous messages are sent.
We also present some experimental results using both models. The results include (1) a verification of the LogGPS model, (2) an example of synchronization analysis using an MPI program and (3) a comparison of the models. The results indicate that the LogGPS model is more accurate than the LogGP model, and analyzing synchronization costs is important when improving parallel program performance.
- 1 K. Al-Tawil and C. A. Moritz. LogGP Quantified: The Case for MPI. In Proc. 7th IEEE International Symp. on High Performance Distributed Computing (HPDC-7), Chicago, IL, August 1998.Google Scholar
Cross Ref
- 2 A. Alexandrov, M. Ionescu, K. Schauser, and C. Scheiman. LogGP: Incorporating Long Messages into the LogP Model - One Step Closer Towards a Realistic Model for Parallel Computation. In Proc. 7th Ann. ACM Symp. on Parallel Algorithms and Architectures (SPAA'95), pages 95-105, Santa Barbara, CA, July 1995. Google ScholarDigital Library
- 3 N. J. Boden, D. Cohen, R. E. Felderman, A. E. Kulawik, C. L. Seitz, J. N. Seizovic, and W.-K. Su. Myrinet-A Gigabit-per-Second Local-Area Network. IEEE Micro, 15(1):29-36, February 1995. Google ScholarDigital Library
- 4 G. Burns, R. Daoud, and J. Vaigl. LAM: An Open Cluster Environment for MPI. In In Proc. of Supercomputing Symposium '94 (SS'94), pages 379-386, Toronto, June 1994. available from http://www.mpi.nd.edu/lam/.Google Scholar
- 5 D. Culler, R. Karp, D. Patterson, A. Sahay, K. Schauser, E. Santos, R. Subramonian, and T. von Eicken. LogP: Towards a realistic model of parallel computation. In Proc. 4th ACM SIGPLAN Symp. on Principles and Practice ofParallel Programming (PPoPP'93), pages 1-12, San Diego, CA, May 1993. Google ScholarDigital Library
- 6 A. C. Dusseau, D. E. Culler, K. E. Schauser, and R. P. Martin. Fast Parallel Sorting Under LogP: Experience with the CM-5. IEEE Transactions on Parallel and Distributed Systems, 7(8):791-805, August 1996. Google ScholarDigital Library
- 7 W. Gropp, E. Lusk, N. Doss, and A. Skjellum. A High-Performance, Portable Implementation of the MPI Message Passing Interface Standard. Parallel Computing, 22(6):789-828, 1996. available from http://www.mcs.anl.gov/mpi/mpich/. Google ScholarDigital Library
- 8 M. Homewood and M. McLaren. Meiko CS-2 interconnect elan - elite design. In Proc. IEEE Hot Interconnects '93 Symp., pages 95-105, August 1993.Google Scholar
- 9 IBM Corp. IBM Parallel Enviroment for AIX, MPL Programming and Subroutine Reference, Document Number GC23-3893-00, 1995.Google Scholar
- 10 IBM Corp. IBM Parallel Enviroment (PE), http://www.rs6000.ibm.com/software/sp products/ pe.html, 2000.Google Scholar
- 11 Y. Kanoh, M. Nakamura, T. Hirose, T. Hosomi, and T. Nakata. User-level Network Interface for a Parallel Computer Cenju-4. Trans. of Information Processing Society of Japan, 41(5):1379-1389, 1999.Google Scholar
- 12 T. Kielmann, H. E. Bal, and K. Verstoep. Fast Measurement ofLogPParameters for Message Passing Platforms. In Proc. 4th Workshop on Runtime Systems for Parallel Programming, pages 1176-1183, May 2000. Google ScholarDigital Library
- 13 Message Passing Interface Forum. MPI: A Message-Passing Interface Standard. International Journal of Supercomputing Applications, 8(3/4), 1994.Google Scholar
- 14 C. A. Moritz. Cost Modeling and Analysis: Towards Optimal Resource Utilization in Parallel Compuer Systems. PhD thesis, Royal Institute of Technology, Stockholm, 1998.Google Scholar
- 15 F. O'Carroll, H. Tezuka, A. Hori, and Y. Ishikawa. The Design and Implementation of Zero Copy MPI Using Commodity Hardware with a High Performance Network. In International Conference on Supercomputing '98 (SC98), pages 243-250, July 1998. available from http://pdswww.rwcp.or.jp/dist/score/. Google ScholarDigital Library
- 16 Parallel Software Contest (PSC95). http://www.info.waseda.ac.jp/muraoka/project/ psc95/, 1995.Google Scholar
- 17 D. Sundaram-Stukel and M. K. Vernon. Predictive Analysis of a Wavefront Application Using LogGP. In Proc. 7th ACM SIGPLAN Symp. on Principles and Practices of Parallel Programming (PPoPP'99), pages 141-150, Atlanta, GA, May 1999. Google ScholarDigital Library
- 18 O. Tatebe. Software, http://phase.etl.go.jp/ %7etatebe/software/index-j.html, 2000.Google Scholar
- 19 T. von Eicken, D. E. Culler, S. C. Goldstein, and K. E. Schauser. Active Messages: A mechanism for integrated communication and computation. In Proc. 19th Ann. International Symp. on Computer Architecture (ISCA), pages 256-266, Gold Coast, May 1992. Google ScholarDigital Library
Index Terms
- LogGPS: a parallel computational model for synchronization analysis
Recommendations
LogGPS: a parallel computational model for synchronization analysis
PPoPP '01: Proceedings of the eighth ACM SIGPLAN symposium on Principles and practices of parallel programmingWe present a new parallel computational model, named LogGPS, which captures synchronization.
The LogGPS model is an extension of the LogGP model, which abstracts communication on parallel platforms. Although the LogGP model captures long messages with ...
The performance and scalability of SHMEM and MPI-2 one-sided routines on a SGI Origin 2000 and a Cray T3E-600: Performances
This paper compares the performance and scalability of SHMEM and MPI-2 one-sided routines on different communication patterns for a SGI Origin 2000 and a Cray T3E-600. The communication tests were chosen to represent commonly used communication patterns ...
OpenMP for Networks of SMPs
In this paper, we present the first system that implements OpenMP on a network of shared-memory multiprocessors. This system enables the programmer to rely on a single, standard, shared-memory API for parallelization within a multiprocessor and between ...
Comments