高性能集群系统分析、设计与应用
|
摘要
高性能集群是一组相互独立的服务器通过高速网络连接构成的高性能计算系统。相对于其它高性能计算机，它的优点是可扩展性好，具有高可用性和很高的性价比。
本文通过对高性能集群技术的分析和对集群互连网络的测试比较，选用SCI可扩展一致性接口技术设计了一个高性能集群，并实现了气象并行计算在高性能集群上的应用。
全文由四部分构成：高性能集群技术、高速互连网络的分析、高性能集群的设计和高性能集群在气象数值预报计算中的应用。
论文首先阐述了高性能计算机的需求和发展现状，研究了高性能集群的体系结构和软、硬件组成，然后详细分析了对集群性能起重要作用的几种高速互连网络的结构和组成，并通过测试对它们的性能指标做了比较。在此基础上，根据气象预报计算的具体需求，采用16节点和SCI网络，设计了高性能集群。同时对该集群网络(SCI)的可用性进行了深入的分析。最后为验证集群的性能，实现了气象数值预报中短期中尺度模式MM5、通用气候模式CCM3和区域气候模式RegCM3在该集群上的调试计算。
综上所述，本文完成了一个高性能集群的详细设计。同时在集群互连技术的分析比较中提供了大量有用的数据，为今后进一步设计大规模的集群奠定了基础。
High performance cluster is a group of single servers that connected together to form a high performance computing system through the high-speed network. Compared with other high performance computers, the great virtue of the cluster is excellent scalability, high reliability and challenging performance value rate.
Based on the analysis of the high performance cluster and the comparison of the cluster's network, the thesis chooses the Scalable Coherent Interface (SCI) to design the cluster, and then realizes the weather parallel application on it.
In the first part, the thesis describes the international trend of high performance cluster and researches the system structure of cluster. Secondly, several high-speed networks of the cluster are analyzed, including the structures and the components of them. Through the analysis, the result of performance comparison test of these networks is provided. According to the need of weather forecast computing, the thesis adopts 16 nodes and SCI to construct the cluster. Finally, the thesis realizes the application of weather numerical value forecast in the cluster, such as Mesoscale Model 5(MM5), Common Climate Model (CCM3) and Regional Climate Model 3(RegCM3).
In conclusion, a high performance cluster system is designed and analyzed, which provides lots of useful data for advanced research.
本文通过对高性能集群技术的分析和对集群互连网络的测试比较，选用SCI可扩展一致性接口技术设计了一个高性能集群，并实现了气象并行计算在高性能集群上的应用。
全文由四部分构成：高性能集群技术、高速互连网络的分析、高性能集群的设计和高性能集群在气象数值预报计算中的应用。
论文首先阐述了高性能计算机的需求和发展现状，研究了高性能集群的体系结构和软、硬件组成，然后详细分析了对集群性能起重要作用的几种高速互连网络的结构和组成，并通过测试对它们的性能指标做了比较。在此基础上，根据气象预报计算的具体需求，采用16节点和SCI网络，设计了高性能集群。同时对该集群网络(SCI)的可用性进行了深入的分析。最后为验证集群的性能，实现了气象数值预报中短期中尺度模式MM5、通用气候模式CCM3和区域气候模式RegCM3在该集群上的调试计算。
综上所述，本文完成了一个高性能集群的详细设计。同时在集群互连技术的分析比较中提供了大量有用的数据，为今后进一步设计大规模的集群奠定了基础。
High performance cluster is a group of single servers that connected together to form a high performance computing system through the high-speed network. Compared with other high performance computers, the great virtue of the cluster is excellent scalability, high reliability and challenging performance value rate.
Based on the analysis of the high performance cluster and the comparison of the cluster's network, the thesis chooses the Scalable Coherent Interface (SCI) to design the cluster, and then realizes the weather parallel application on it.
In the first part, the thesis describes the international trend of high performance cluster and researches the system structure of cluster. Secondly, several high-speed networks of the cluster are analyzed, including the structures and the components of them. Through the analysis, the result of performance comparison test of these networks is provided. According to the need of weather forecast computing, the thesis adopts 16 nodes and SCI to construct the cluster. Finally, the thesis realizes the application of weather numerical value forecast in the cluster, such as Mesoscale Model 5(MM5), Common Climate Model (CCM3) and Regional Climate Model 3(RegCM3).
In conclusion, a high performance cluster system is designed and analyzed, which provides lots of useful data for advanced research.
引文
[1] 张继超、舒继武、郑纬民、常迪，《基于Myrinet／GM多通道通信》，《软件学报》，2003年第14卷第2期。
[2] 都志辉，《高性能计算并行编程技术—MPI并行程序设计》，清华大学出版社，2001年。
[3] 陈志辉、马捷、陈国良、高帆，《用户级通信协议BCL-3对IP协议支持的研究》，《软件学报》，2003年第14卷第9期。
[4] 郑纬民，石威，汪东升，《高性能集群计算：结构与系统(第一卷)》，电子工业出版社，2001年6月。
[5] 车静光，《微机集群组建、优化和管理》，机械工业出版社，2004年1月。
[6] Derek，《LINUX RAID管理(影印版)》，清华大学出版社，2003年10月。
[7] 冯百明，《基于分区的自动并行化程序重构技术研究》，西北工业大学出版社，1999年。
[8] Rich Seifert, Gigabit Ethernet: technology and applications for high-speed LANs, 机械工业出版社，2000年10月。
[9] Weikuan Yu, Darius Buntinas, Dhabaleswar K. Panda, High Performance and Reliable NIC-Based Multicast over Myrinet/GM-2, Dept. of Computer and Information Science, The Ohio State University, 2003.
[10] Jenwei Hsieh, Tau Leng, Victor Mashayekhi, Reza Rooholamini, Architectural and Performance Evaluation of GigaNet and Myrinet Interconnects on Clusters of Small-Scale SMP Servers. Enterprise System Group, Dell Computer Corporation, 2003.
[11] Andrew Begel, Philip Buonadonna, David E. Culler, David Gay, An Analysis of Ⅵ ArchitecCure Primitives in Support of Parallel and Distributed Communication, University of California, Berkeley, 2003.
[12] Hermann Hellwagner, Markus Lachowitz, Matthias Ohlenroth, Exploring
the Performance of Ⅵ Architecture Commnnication Features in the GigaNet Cluster LAN, Institute of IT, University Klagenfurt, 2001.
[13] R. S. Madukkarumukumana, C. Pu and H. V. Shah, Harnessing User-Level Networking Architectures for Distributed Object Computing over High-Speed Networks, Prec. of the 2nd USENIX Windows NT Symposium, Aug. 1998, pp. 127-135.
[14] Fabrizio Petrini, Eitan Franchtenberg, Adolfy Hoisie, Performance Evaluation of the Quadrics Interconnection Network, Cluster Computing 6, 2003, pp. 125-136.
[15] Petrini F, Feng WC, Hoisie A, Cell S, Frachtenberg E, The quadrics network(QsNet): high-performance clustering technology. In: Proceedings of the 9th IEEE Hot Interconnects(HotI 2001), IEEE Computer Society Press, 2001, pp. 125-133.
[16] Cesar A. F. De Rose, Reynaldo Novaes, Tiago Ferreto, Fabio A. D. de Oliveira, Marcos E. Barreto, Rafael B. Avila, Philippe O. A. Navaux, Hans-Ulrich Heiss, The Scalable Coherent Interface(SCI) as an Alternative for Cluster Interconnection, UFRGS. Brazil, 2001.
[17] Stein Jorgen Ryan, Haakon Bryhni, SCI for Local Area Networks, University of Oslo, Jan. 1998.
[18] Maximilian Ibel, Klaus E. Schauser, Chris J. Scheiman, Manfred Weis, High-Performance Cluster Computing Using SCI, Dep. Computer Science University of California, Santa Barbara, 1999.
[19] Ch. Kurmann, T. Stricker, A Comparison of two Gigebit SAN/LAN technologies: Scalable Coherent Interface versus Myrinet, Laboratory for Computer Systems Swiss Institute of Technology(ETH), 2000.
[20] S. Millich, A. George, S. Oral, A Comparative Throughput Anaiysis of Scalable Coherent Interface and Myrinet, HCS Research Lab, ECE Dept., University of Florida, 2002.
[21] D. Gonzalez, A. George, M. Chidester, Performanco Modeling and
Evaluation of Topologies for Low-Latency SCI Systems, Microprocessor and Microsystems, Vol. 25, No. 7, Oct. 2001, pp. 343-356.
[22] J. Hsieh, T. Leng, V. Mashayekhi, R. Rooholamini, Architectural and Performance Evaluation of GigaNet and Myrinet Interconnects on Clusters of Small-Scale SMP Servers, Proc. of IEEE Supercomputing (SC' 2000), Dallas, USA, 2000.
[23] Arun Lyengar, Jim Challenger, Improving Web Awever PerFormance by Caching Dynamic Data, Proceedings of the Usenix Symposium on Internet Technologies and Systems, Dec. 1997.
[24] IBM Corporation, DataLinks: Managing External Data with DB2 Universal Database, White Paper, Feb. 2000.
[25] Rafat Mukherjee, Real-Time Virtual Shared Disk: Enabling Multimedia on Clusters, Proceedings of the PDPTA-98 Conference, July 1998.
[26] R. Dimitrov, A. Skjellum, Impact of Latency on Applications Performance, Proceedings of the Fourth MPI Developer's and User's Conference, Cornell University, Apr. 2000.
[27] GigaNet. Inc, GigaNet cLAN Family of Products, http://www. giganet. com/, 2003.
[28] Tom Clark, IP SANs: Guide to iSCSI, iFCP and FCIP Protocols for SAN, pearson Education, 2003.
[29] Erik B. Kluzek, Jerry Olson, James M. Rosinski, John E. Truesdale, Mariana Vertenstein, User's Guide to NCAR CCM3. 6, NCAR, 2003.
[30] Miramare Trieste, RegCM Version 3.0 User's Guide, Physics Weather and Climate Group International Center of Theoretical Physics, 2003.
[31] 冯圣中，《并行计算基础知识》，中国科学院计算技术研究所，2003年。
[32] 张云泉，《2003年中国高性能计算机性能TOP100排行榜分析》，中国数学软件分会，2003年。
[2] 都志辉，《高性能计算并行编程技术—MPI并行程序设计》，清华大学出版社，2001年。
[3] 陈志辉、马捷、陈国良、高帆，《用户级通信协议BCL-3对IP协议支持的研究》，《软件学报》，2003年第14卷第9期。
[4] 郑纬民，石威，汪东升，《高性能集群计算：结构与系统(第一卷)》，电子工业出版社，2001年6月。
[5] 车静光，《微机集群组建、优化和管理》，机械工业出版社，2004年1月。
[6] Derek，《LINUX RAID管理(影印版)》，清华大学出版社，2003年10月。
[7] 冯百明，《基于分区的自动并行化程序重构技术研究》，西北工业大学出版社，1999年。
[8] Rich Seifert, Gigabit Ethernet: technology and applications for high-speed LANs, 机械工业出版社，2000年10月。
[9] Weikuan Yu, Darius Buntinas, Dhabaleswar K. Panda, High Performance and Reliable NIC-Based Multicast over Myrinet/GM-2, Dept. of Computer and Information Science, The Ohio State University, 2003.
[10] Jenwei Hsieh, Tau Leng, Victor Mashayekhi, Reza Rooholamini, Architectural and Performance Evaluation of GigaNet and Myrinet Interconnects on Clusters of Small-Scale SMP Servers. Enterprise System Group, Dell Computer Corporation, 2003.
[11] Andrew Begel, Philip Buonadonna, David E. Culler, David Gay, An Analysis of Ⅵ ArchitecCure Primitives in Support of Parallel and Distributed Communication, University of California, Berkeley, 2003.
[12] Hermann Hellwagner, Markus Lachowitz, Matthias Ohlenroth, Exploring
the Performance of Ⅵ Architecture Commnnication Features in the GigaNet Cluster LAN, Institute of IT, University Klagenfurt, 2001.
[13] R. S. Madukkarumukumana, C. Pu and H. V. Shah, Harnessing User-Level Networking Architectures for Distributed Object Computing over High-Speed Networks, Prec. of the 2nd USENIX Windows NT Symposium, Aug. 1998, pp. 127-135.
[14] Fabrizio Petrini, Eitan Franchtenberg, Adolfy Hoisie, Performance Evaluation of the Quadrics Interconnection Network, Cluster Computing 6, 2003, pp. 125-136.
[15] Petrini F, Feng WC, Hoisie A, Cell S, Frachtenberg E, The quadrics network(QsNet): high-performance clustering technology. In: Proceedings of the 9th IEEE Hot Interconnects(HotI 2001), IEEE Computer Society Press, 2001, pp. 125-133.
[16] Cesar A. F. De Rose, Reynaldo Novaes, Tiago Ferreto, Fabio A. D. de Oliveira, Marcos E. Barreto, Rafael B. Avila, Philippe O. A. Navaux, Hans-Ulrich Heiss, The Scalable Coherent Interface(SCI) as an Alternative for Cluster Interconnection, UFRGS. Brazil, 2001.
[17] Stein Jorgen Ryan, Haakon Bryhni, SCI for Local Area Networks, University of Oslo, Jan. 1998.
[18] Maximilian Ibel, Klaus E. Schauser, Chris J. Scheiman, Manfred Weis, High-Performance Cluster Computing Using SCI, Dep. Computer Science University of California, Santa Barbara, 1999.
[19] Ch. Kurmann, T. Stricker, A Comparison of two Gigebit SAN/LAN technologies: Scalable Coherent Interface versus Myrinet, Laboratory for Computer Systems Swiss Institute of Technology(ETH), 2000.
[20] S. Millich, A. George, S. Oral, A Comparative Throughput Anaiysis of Scalable Coherent Interface and Myrinet, HCS Research Lab, ECE Dept., University of Florida, 2002.
[21] D. Gonzalez, A. George, M. Chidester, Performanco Modeling and
Evaluation of Topologies for Low-Latency SCI Systems, Microprocessor and Microsystems, Vol. 25, No. 7, Oct. 2001, pp. 343-356.
[22] J. Hsieh, T. Leng, V. Mashayekhi, R. Rooholamini, Architectural and Performance Evaluation of GigaNet and Myrinet Interconnects on Clusters of Small-Scale SMP Servers, Proc. of IEEE Supercomputing (SC' 2000), Dallas, USA, 2000.
[23] Arun Lyengar, Jim Challenger, Improving Web Awever PerFormance by Caching Dynamic Data, Proceedings of the Usenix Symposium on Internet Technologies and Systems, Dec. 1997.
[24] IBM Corporation, DataLinks: Managing External Data with DB2 Universal Database, White Paper, Feb. 2000.
[25] Rafat Mukherjee, Real-Time Virtual Shared Disk: Enabling Multimedia on Clusters, Proceedings of the PDPTA-98 Conference, July 1998.
[26] R. Dimitrov, A. Skjellum, Impact of Latency on Applications Performance, Proceedings of the Fourth MPI Developer's and User's Conference, Cornell University, Apr. 2000.
[27] GigaNet. Inc, GigaNet cLAN Family of Products, http://www. giganet. com/, 2003.
[28] Tom Clark, IP SANs: Guide to iSCSI, iFCP and FCIP Protocols for SAN, pearson Education, 2003.
[29] Erik B. Kluzek, Jerry Olson, James M. Rosinski, John E. Truesdale, Mariana Vertenstein, User's Guide to NCAR CCM3. 6, NCAR, 2003.
[30] Miramare Trieste, RegCM Version 3.0 User's Guide, Physics Weather and Climate Group International Center of Theoretical Physics, 2003.
[31] 冯圣中，《并行计算基础知识》，中国科学院计算技术研究所，2003年。
[32] 张云泉，《2003年中国高性能计算机性能TOP100排行榜分析》，中国数学软件分会，2003年。
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |