Abstract
We have developed a new communication library, called PM, for the Myrinet gigabit LAN card, that has a dedicated processor and onboard memory to handle communication protocols. To obtain high performance communication and support multi-user environments, we have co-designed PM, an operating system implemented as a daemon process, and the run-time routine for a programming language. Several unique features, e.g., network context switching and a Modified ACK/NACK flow control algorithm, have been developed for PM.
The PM library has been implemented on two types of clusters: Sun SPARCstation model 20/71 workstations and Intel Pentium based PCs. PM on the Sun workstations has a round trip time of 20 μseconds for a user-level 8 byte message and a bandwidth of 38.6 Mbytes/second for an 8 Kbyte message. The result of a NAS parallel benchmark shows that a Sparc 20 workstation cluster achieves almost the same performance as a Cray T3D.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
N. J. Boden, D. Cohen, R. E. Felderman, A. E. Kulawik, C. L. Seitz, J. N. Seizovic and Wen-King Su. “Myrinet — A Gigabit-per-Second Local-Area Network”. IEEE MICRO, Vol. 15, No. 1, pp. 29–36, February 1995.
http://now.cs.berkeley.edu/AM/lam_release.html.
Scott Pakin, Mario Lauria and Andrew Chein. “High Performance Messaging on Workstations: Illinois Fast Messages (FM) for Myrinet”. In Proceedings of Supercomputing '95, San Diego, California, 1995.
Thorston von Eicken, Anindya Basu, and Werner Vogels. U-Net: A User Level Network Interface for Parallel and Distributed Computing. In Fifteenth ACM Symposium on Operating Systems Principles, pp. 40–53, 1995.
Y. Ishikawa, A. Hori, H. Tezuka, M. Matsuda, H. Konaka, M. Maeda and T. Tomokiyo, and J. Nolte. MPC++. In Gregory V. Wilson and Paul Lu, editors, Parallel Programming Using C++, pp. 429–464. MIT Press, 1996.
Y. Ishikawa. Multi Thread Template Library — MPC++ Version 2.0 Level 0 Document —. Technical Report TR-96012, RWC, September 1996. Visit http://www.rwcp.or.jp/lab/mpslab/mpc++/mpc++.html.
A. Hori, T. Yokota, Y. Ishikawa, S. Sakai, H. Konaka, M. Maeda, T. Tomokiyo, J. Nolte, H. Matsuoka, K. Okamoto, and H. Hirono. Time Space Sharing Scheduling and Architectural Support. In D. G. Feitelson and L. Rudolph, editors, Job Scheduling Strategies for Parallel Processing, Vol. 949 of Lecture Notes in Computer Science. Springer-Verlag, April 1995.
A. Hori, H. Tezuka, Y. Ishikawa, N. Soda, H. Konaka, and M. Maeda. Implementation of Gang-Scheduling on Workstation Cluster. In D. G. Feitelson and L. Rudolph, editors, IPPS'96 Workshop on Job Scheduling Strategies for Parallel Processing, Vol. 1162 of Lecture Notes in Computer Science, pp. 76–83. Springer-Verlag, April 1996.
T. von Eicken, V. Avula, A. Basu and V. Buch. “Low-Latency Communication over ATM Networks using Active Messages”. In Proceedings of Hot Interconnects II, 1994 Palo Alto, August 1994.
http://www.myri.com.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Tezuka, H., Hori, A., Ishikawa, Y., Sato, M. (1997). PM: An operating system coordinated high performance communication library. In: Hertzberger, B., Sloot, P. (eds) High-Performance Computing and Networking. HPCN-Europe 1997. Lecture Notes in Computer Science, vol 1225. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0031642
Download citation
DOI: https://doi.org/10.1007/BFb0031642
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-62898-9
Online ISBN: 978-3-540-69041-2
eBook Packages: Springer Book Archive