【作者】 张鲁飞； 程华；

【Author】 Zhang Lu -fei~1,Cheng Hua~(1,2) (1.Jiangnan Institute of Computing Technology,Wuxi 214083,China;2.Institute of Computing Technology, Chinese Academy of Sciences/Graduate School of the Chinese Academy of Sciences,Beijing 100080,China.)

【机构】 江南计算技术研究所； 中国科学院计算技术研究所； 中国科学院研究生院；

【摘要】 作为绿色计算的核心,绿色操作系统需要适应计算及体系结构和运行于其上的应用程序新的发展,从而达到"节能、环保"的。本文将首先研究异构多核平台、低功耗集群和云计算模式中绿色操作系统的作用、功能特点及实现方法,然后介绍操作系统中电源管理研究的新进展,最后对当前面临的主要问题和发展趋势进行了总结和展望。更多还原

【Abstract】 As the scale of High Performance Computer is continually expanding,the increasing system errors become a serious threaten to its performance.This paper introduces the concept of availability of HPC systems and explores the root cause of failures.Then some traditional techniques for fault tolerance and their challenges are outlined and some novel high availability techniques and their prototypes are introduced.Finally,several challenges and open issues in high availability techniques in HPC systems are summarized.更多还原