期刊文献+

提高工作站机群系统通信性能方法的研究 被引量:2

HOW TO IMPROVE COMMUNICATION PERFORMANCE FOR NETWORK OF WORKSTATIONS
下载PDF
导出
摘要 工作站机群系统是目前并行处理技术的研究热点之一,而如何提高其网络通信的性能又是发展机群系统的关键。本文在分析影响网络系统性能因素的基础上,从软、硬件两方面着重讨论了高速网络技术。 Network of Workstations is becoming one of the leading technique in parallel processing, however, its poor communication performance seriously affects the speedup and efficiency of parallel computation. In this paper, we analyze four elements which dominate the cost of message-passing, then discuss several methods to minimize the communication overhead, such as high speed network technology, reduced communication protocol and Active Message. At last, we present two systems to show that these methods can make the networks more efficient.
出处 《小型微型计算机系统》 EI CSCD 北大核心 1997年第6期8-13,共6页 Journal of Chinese Computer Systems
基金 国家863计划资助
关键词 工作站机群系统 高速网络 精简通信协议 Network of workstations, High speed network, Reduced communication protocol, Active message
  • 相关文献

参考文献1

  • 1周明天,TCP/IP网络原理与技术,1993年

同被引文献19

  • 1Bruck J,Dolev D,Ho Ching-Tien. Efficient Message Passing Interface (MPI) for Parallel Computing on Clusters of Workstation. Journal of Parallel and Distributed Computing, 1997,40:19~34
  • 2Pakin S,Karamcheti V, Chien A A. Fast Message(FM): Efficient, Portable Communication for Workstation Clusters and Massively-Parallel Processors. IEEE Concurrency, 1997,5(2): 60~73
  • 3Lauria M,Chien A. MPI-FM: High Performance MPI on Workstation Cluster. Journal of Parallel and Distributed Computing,1997,40:4~18
  • 4Lauria M,Pakin S,Chien A. Efficient Layering for High Speed Communication: the MPI over Fast Message (FM) Experience.Cluster Computing,1999,2(2): 107~116
  • 5Jacunski M,et al, Low Latency Message-Passing for Reflective Memory Networks. In: CANPC'99, LNCS 1602, 1999. 211~224
  • 6von Eicken T,et al. Active Messages: a mechanism for integrated communication and computation. In:Proc. of the Intl. Symposium on Computer Architecture, 1992. 256~266
  • 7Rodrigues S,Anderson T,Culler D. High-performance local-area communication using Fast Socket. In:Proc. of the USENIX 1997Technical Conf. San Diego, California, USENIX Association,1997
  • 8Yocum K G,et al. Cut-through delivery in Trapeze: an exercise in low-latency messaging. In HPDC-6, Portland, Oregon, August 1997
  • 9von Eicken T,et al. U-Net: A user-level network interface for parallel and distributed computing. In: Proc. of the 15th ACM Symposium on Operating Systems Principles. Dec. 1995.40~ 53
  • 10Dubnicki C,et al. VMMC-2: efficient support for reliable, con nection-oriented communication. In:Proc. of Hot Interconnects V. IEEE, Aug. 1997

引证文献2

二级引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部