期刊文献+

基于多虚空间多重映射技术的并行操作系统 被引量:3

A Parallel Operating System Based on Multi-Virtual-Space and Multi-Mapping Technology
下载PDF
导出
摘要 高性能计算机系统的可扩展性是系统设计的一大难题 ,NUMA(non-uniform memory architecture)结构正是为了解决共享存储体系的可扩展性问题而提出来的 .研究和实践表明 ,整机系统的可扩展性与操作系统的结构有着密切的关系 .典型的多处理机操作系统通常采用两种结构 ,基于共享的单一核心结构以及基于消息的多核心结构 .通过分析得出结论认为 ,这两种结构都不能很好地适应可扩展并行机尤其是 NUMA结构并行机的需求 .针对存在的问题 ,提出了新的结构设计思想 :多虚空间多重映射与主动消息相结合 .测试和运行结果显示 。 The scalability of system is a puzzle to design high performance computers. NUMA (non -uniform memory architecture) architecture is proposed for scalable system underlying a shared memory pattern. Research and practice results show that the scalability of system tightly correlates with that of the operating system. Usually, two structures of operating system are employed in multiprocessor systems, monolithic based on shared memory and multi kernel based on message passing. However, both of them cannot fully fit scalable parallel computers, especially MPP systems with NUMA architecture. In this paper, the problems in two structures are analyzed, a new structure, multi virtual space and multi mapping with active message, is proposed. The test and practice results show that the new structure is successful in the scalability of system. This paper introduces detailedly the implementation of the structure in a MPP system with NUMA architecture.
出处 《软件学报》 EI CSCD 北大核心 2001年第10期1562-1568,共7页 Journal of Software
关键词 并行操作系统 多重映射技术 可扩展性 计算机系统 operating system scalability programmability NUMA (non -uniform memory architecture) architecture massively parallel processing
  • 相关文献

参考文献7

  • 1[1]Erik, Hager, Greg, Papadopoulos. Parallel computing in the commercial marketplace: research and innovation at work. Proceedings of the IEEE, 1999,87(3):405~411.
  • 2[2]Nukherjee, B. A survey of multiprocessor operating system kernels. Technical Report, GITCC-92/05, Georgia Institute of Technology, 1993.
  • 3[3]Chapin, J., Mendel, Rosenblum, Scott, Devine, et al. Hive: fault containment for shared-memory multiprocessors. In: Proceedings of the 15th ACM Symposium on Operating Systems Principles. New York: ACM Press, 1995. http://sdg.lcs.mit.edu/~jchapin/ publications.html.
  • 4[4]Kaushik, Ghosh, Christie, A.J. Communication across fault-containment firewalls on the SGI origin. In: Proceedings of the International Symposium on High-Performance Computer Architecture. Los Alamitos, CA: IEEE Computer Society Press, 1998.
  • 5[5]Ben, Verghese, Scott, Devine, Anoop, Gupta, et al. Operating system support for improving data locality on CC-NUMA compute servers. In: Berman, A.M., ed. Proceedings of the 7th International Conference on Architectural Support for Programming Languages and Operating Systems. New York: ACM Press, 1996. 279~289.
  • 6[6]Bruce, Walker, Douglas, Steel. Implementing a full single image Unixware cluster: Middleware vs Underware. In: Proceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications. Los Alamitos, CA: IEEE Computer Society Press, 1999. http://www.dgs.monsh.edu.au/~rajkumar/pdpta99.
  • 7[7]Von Eicken, T., Culler, D.E., Goldstein, S.C., et al. Active message: a mechanism for integrated communication and computation. In: Proceedings of the 19th Annual International Symposium on Computer Architecture. Los Alamitos, CA: IEEE Computer Society Press, 1992.

同被引文献60

引证文献3

二级引证文献10

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