Skip to main content

Evaluation of real-time fiber communications for parallel collective operations

  • Conference paper
  • First Online:
Parallel and Distributed Processing (IPPS 1999)

Part of the book series: Lecture Notes in Computer Science ((LNCS,volume 1586))

Included in the following conference series:

Abstract

Real-Time Fiber Communications (RTFC) is a gigabit speed network that has been designed for damage tolerant local area networks. In addition to its damage tolerant characteristics, it has several features that make it attractive as a possible interconnection technology for parallel applications in a cluster of workstations. These characteristics include support for broadcast and multicast messaging, memory cache in the network interface card, and support for very fine grain writes to the network cache. Broadcast data is captured in network cache of all workstations in the network providing a distributed shared memory capability. In this paper, RTFC is introduced. The performance of standard MPI collective communications using TCP protocols over RTFC are evaluated and compared experimentally with that of Fast Ethernet. It is found that the MPI message passing libraries over traditional TCP protocols over RTFC perform well with respect to Fast Ethernet. Also, a new approach that uses direct network cache movement of buffers for collective operations is evaluated. It is found that execution time for parallel collective communications may be improved via effective use of network cache.

This work was supported in part by NSF Research Instrumentation Grant #9617384 and by equipment support from Belobox Systems, Inc., Irvine, California.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Institutional subscriptions

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Bahaa-El-Din, W. H., and Liu, M. T. Register insertion: A protocol for the next generation of ring local-area networks. Computer Networks and ISDN Systems 24 (1992), 349–366.

    Article  MATH  Google Scholar 

  2. Belobox Systems Inc., AmpDC Advanced Multiprocessor Distributed Computing Environment, 1997. http://www.belobox.com/ampdc.html.

    Google Scholar 

  3. Dunigan, T. H., and A. Hall, K. PVM and IP multicast. Tech. Rep. ORNL Report ORNL/TM-13030, Oak Ridge National Laboratory, December 1996.

    Google Scholar 

  4. Gropp, W., Lusk, E., Doss, N., and Skjellum, A. A High-Performance, Portable Implementation of the MPI Message Passing Interface Standard. Tech. Rep. Preprint MCS-P567-0296, Argonne National Laboratory, March 1996.

    Google Scholar 

  5. Gwinn, J. RTFC Principles of Operation, RTFC-2001. Tech. rep., Belobox Systems Inc., Irvine, California, January 2 1998.

    Google Scholar 

  6. Martin, R. P. HPAM: An active message layer for a network of HP workstations. In Hot Interconnects II (August 1994).

    Google Scholar 

  7. Nupairoj, N., and Ni, L. Benchmarking of multicast communication services. Tech. Rep. Technical Report MSU-CPS-ACS-103, Department of Computer Science, Michigan State University, April 1995.

    Google Scholar 

  8. Rajagopal, P. Study of the network cache architecture as a platform for parallel applications. Master’s thesis, Vanderbilt University, August 1998.

    Google Scholar 

  9. RTFC-team. RTFC Micropacket Data Link Layer, RTFC-2002. Tech. rep., Belobox Systems Inc., Irvine, California, October 1997.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Editor information

José Rolim Frank Mueller Albert Y. Zomaya Fikret Ercal Stephan Olariu Binoy Ravindran Jan Gustafsson Hiroaki Takada Ron Olsson Laxmikant V. Kale Pete Beckman Matthew Haines Hossam ElGindy Denis Caromel Serge Chaumette Geoffrey Fox Yi Pan Keqin Li Tao Yang G. Chiola G. Conte L. V. Mancini Domenique Méry Beverly Sanders Devesh Bhatt Viktor Prasanna

Rights and permissions

Reprints and permissions

Copyright information

© 1999 Springer-Verlag

About this paper

Cite this paper

Rajagopal, P., Apon, A.W. (1999). Evaluation of real-time fiber communications for parallel collective operations. In: Rolim, J., et al. Parallel and Distributed Processing. IPPS 1999. Lecture Notes in Computer Science, vol 1586. Springer, Berlin, Heidelberg . https://doi.org/10.1007/BFb0097912

Download citation

  • DOI: https://doi.org/10.1007/BFb0097912

  • Published:

  • Publisher Name: Springer, Berlin, Heidelberg

  • Print ISBN: 978-3-540-65831-3

  • Online ISBN: 978-3-540-48932-0

  • eBook Packages: Springer Book Archive

Publish with us

Policies and ethics