High-Performance Networking for Software DSMs
Resumo
Several messaging software architectures (MSAs) have been proposed and implemented for high-performance local-area networks (LANs). Several of these MSAs have been success ful at providing low latency and high bandwidth to user-level processes that communicate via explicit message passing. In this paper we claim that these MSAs are suboptimal for page-based software distributed shared-memory systems (software DSMs), as they do not consider the specific characteristics of these systems. We support our claim by studying the communication behavior of several applications running on top of the TreadMarks system and by showing that no previously-proposed architecture is ideal for the observed behavior. Finally, we propose a novel MSA for the Myrinet LAN that is tailored to software DSMs. This new design includes isolated features from some other MSAs, offering reliable message delivery, optimizations for both short and long messages, and several options for message arrival notification. In addition, our proposed MSA relies on a communication model that simplifies buffor management, while reducing latency response for request-reply operations.
Referências
J. K. Bennett, J. B. Carter, and W. Zwaenepoel. Munin: Distributed Shared Memory Based on Type-Specific Memory Coherence. In Proceedings of the 2nd PPoPP, pages 168-176, March 1990.
R. Bhoedjang, T. Ruhl, and H. Bal. Design Issues for User-Level Network Interface Protocols on Myrinet. Technical report, Dept of Mathematics and Computer Science, Vrije Universiteit. 1998. To appear in IEEE Computer.
N. Boden, D. Cohen, R. Felderman, A. Kulawik, C. Seitz, J. Seizovic, and W. Su. Myrinet: A Gigabit-per-Second Local Area Network. IEEE MICRO. 15(1):19-36. February 1995.
C. Dubnicki, A. Bilas, IX. Li, and J. Philbin. Design and Implementation of Virtual Memory-Mapped Communication on Myrinet. In Proceedings of the 1997 IPPS, pages 388-396, April 1997.
T. Eicken, D. Culler, S. Goldstein, and K. Schauser. Active Messages: A Mechanism for Integrated Communication and Computation. In Proceedings of the 19th ISCA, pages 256- 266, May 1992.
G. Henley, N. Doss, T. McMahon, and A. Skjellum. BDM: A Multiprotocol Myrinet Control Program and Host Application Programmer Interface. Technical Report MSSU-EIRS-ERC-97-3, Mississippi State University, May 1997.
H.Tezuka, A. Hori, Y. Ishikawa, and M. Sato. PM: A Operating System Coordinated High Performance Communication Library. In High-Performance Computing and Networking '97. volume 1225, pages 708-717, April 1997.
V. Karamcheti and A. Chien. Software Overhead in Messaging Layers: Where Does the Time Go? In Proceedings of ASPLOS-IV, 1994.
P. Keleher, S. Dwarkadas, A.L. Cox, and W. Zwaenepoel. TreadMarks: Distributed Shared Memory on Standard Workstations and Operating Systems. In Proceedings of the 1994 Winter Useniz Conference, Jan 1994.
Andrew J. Gallatin Kenneth G. Yocum, Jeffrey S. Chase and Alvin R. Lebeck. Cut-Through Delivery in Trapeze: An Exercise in Low Latency Messaging. In Proceedings of HPDC, August 1997.
L. Monnerat and R. Bianchini. Efficiently Adapting to Sharing Patterns in Software DSMs. In Proceedings of the 4th HPCA, Feb 1998.
Myricom. Myrinet Specifications. http://www.myri.com/myricom/document.html, 1995.
S. Pakin, M. Lauria, and A. Chien. High Performance Messaging on Workstations: Illinois Fast Messages (FM) for Myrinet. In Proceedings of Supercomputing 95, San Diego. CA. 1995.
L. Prylli and B. Tourancheau. BIP: A New Protocol Designed for High-Performance Networking on Myrinet. In Proceedings of IPPS/SPDP98, 1998.
C. B. Seidel, R. Bianchini, and C. L. Amorim. The Affinity Entry Consistency Protocol. In Proceedings of the 1997 ICPP, Aug 1997.
Y. Zhou, L. Iftode, and K. Li. Performance Evaluation of Two Home-Based Lazy Release Consistency Protocols for Shared Memory Virtual Memory Systems. In Proceedings of the 2nd OSDI, October 1996.
