期刊文献+

一种适用于机群系统的用户层消息传递机制 被引量:4

A Fast Message Passing Mechanism at User-Level for Cluster Computing
下载PDF
导出
摘要 网络通信中的软件开销是目前影响机群系统性能的瓶颈 ,为了提高机群系统的通信性能 ,提出了一种用于机群系统的用户层快速消息传递机制 UL FM(user- level fast message) ,并在 TH- GBNet上加以实现 .ULFM以通信缓冲区的管理为核心设计了一种易于实现的用户层网络访问接口 ,通过综合采用用户层网络接口访问、零拷贝数据传递、精简通信协议等技术 ,可以有效地降低机群系统的通信开销 ,为应用程序提供实际的低延迟、高带宽的通信性能 .在 TH- GBNet的实现中 ,UL FM节点间 16字节数据包的单向延迟减小到仅为 6 μs,2 K字节数据包的带宽达 4 0 MB/ The software overhead in IN (interconnection network) communication has become currently the bottleneck of cluster system performance. In this paper, a method is proposed to reduce the communication software overhead by providing a fast message passing mechanism at user level——ULFM (user level fast message), that has been implemented in the IN TH\|GBNet of cluster system. In ULFM, a new type of network access interface is provided to form a communication buffer centric application access to the cluster system IN. Several techniques, such as user level network access, zero copy message passing and simplified communication protocol, are integrated together for ULFM, that can effectively reduce the communication software overhead and achieve high communication performance with low latency and high bandwidth for application programs. Experimental results show that the latency at application level between different nodes for 16 byte message can be reduced to only 6μs, and the bandwidth can reach to 40MB/s for 2K byte length message.
出处 《软件学报》 EI CSCD 北大核心 2001年第5期689-697,共9页 Journal of Software
基金 国家自然科学基金资助项目 (6 99730 2 4)&&
关键词 机群系统 用户层 通信协议 消息传递 网络通信 软件开销 计算机网络 workstation cluster user level communication protocol message passing overhead
  • 相关文献

参考文献4

二级参考文献1

  • 1Dong C L,Proc of APDC’97,1997年,307页

共引文献12

同被引文献27

  • 1王正华,赵文涛.二维不可压流程序在PVM环境下的并行计算[J].空气动力学学报,2002,20(z1):59-63. 被引量:5
  • 2Gopalak rishnan R, Parulkar G M. Efficient user-space protocol implementations with QoS guarantees using real-time upcall[J]. IEEE/ACM Trans Networking, 1998, 6(4): 374-388.
  • 3Boden N J, Cohen D. Myrinet:A Gigabit-per-second loca.
  • 4Christophe A P,Bruce A F.Distributed computing with personal computers[J].AIChE Journal,1996,42(1):290-295.
  • 5High K A,Laroche R D.Parallel nonlinear optimization techniques for chemical process design problems[J].Computers Chem Engng,1995,19 (6-7):807-825.
  • 6Jiawei Han,Micheline Kamber.Data mining:Concepts and techniques[J].Computers Engineering,1996,34(1):120-125.
  • 7YOOK J K,TILBURY D M,SOPARKAR N R.Trading computation for bandwidth: reducing communication in distributed control systems using state estimators[J].IEEE Transactions on Control Systems Technology,2002,10(4):503-518.
  • 8AMAMIYA M,TANIGUCHI H,MATSUZAKI T.An architecture of fusing communication and execution for global distributed processing[J].Parallel Processing Letters,2001,11(1):7-24.
  • 9ANDREWS D,AUSTIN P,COSTELLO P,et al.Interprocess communications in the AN/BSY-2 distributed computer system: a case study[J].Journal of Systems and Software,2002,61(3):233-242.
  • 10SEYEDI A,SAULNIER G J.A distributed algorithm for dynamic sub-channel assignment in a multi-user OFDM communication system[A].In:IEEE Workshop on Statistical Signal Processing Proceedings[C].[s.l.],2001.

引证文献4

二级引证文献7

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