期刊文献+

具有节点亲近能力的NUMA调度算法

NUMA Scheduling Algorithm Based on Affinitive Node
下载PDF
导出
摘要 非一致存储访问(Non-UniformMemoryAccess,NUMA)是目前高性能服务器的主流体系结构之一,传统操作系统的调度器由于无法感知复杂NUMA系统的拓扑结构,导致较大的远程节点数据访问开销。该文在深入分析O(1)调度算法对NUMA支持的基础上,提出一种基于NUMA拓扑结构的分级调度算法。实验证明,该算法较好地实现了具有节点亲近能力的NUMA调度,提高了数据访问的局部性,优化了系统性能。 Nowadays, non-uniform memory access is one of mainstream architectures in high performance server, and traditional operation system's scheduling incurs lots of data access cost to remote node because of lack of topology information for NUMA. This paper presents a hierarchical scheduling algorithm based on NUMA topology after analyzing the 0(1) scheduling algorithm's support for NUMA. The result shows that this algorithm realizes NUMA affinity scheduling, improves data access locality and system performance.
出处 《计算机工程》 EI CAS CSCD 北大核心 2006年第1期99-101,156,共4页 Computer Engineering
基金 国家"863"计划基金资助项目(2002AA1Z2101 2003AA1Z2060)
关键词 非一致存储访问 操作系统 分级调度 节点亲近能力 NUMA Operation system Hierarchical scheduling Affinifive node
  • 相关文献

参考文献5

  • 1Laudon J, Lenoski D. The SGI Origin: A ccNUMA Highly Scaleble Server. Proceedings of the 24^nd International Symposium on Computer Architecture, 1997.
  • 2Lenoski D, Laudon J, Gharachorloo K, et al. The Stanford DASH Multiprocessor[J]. IEEE Computer, 1992, 25(3):63-79.
  • 3Agarwal A, Bianchini R, Chaiken D, et al. The MIT Alewife Machine:Architecture and Performance. Proceedings of the 22^nd Annual International Symposiumon Computer Architecture, 1995:2-13.
  • 4Dobson M, Gaughen R Hohnbaum M. Linux Support for NUMA Hardware. Ottawa, Ontario, Canada: Proceedings of the Linux Symposium, 2003.
  • 5Focht E. Node Affine NUMA Scheduler[Z]. http://home.arcor.de/efocht/ shced/.

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