高性能计算系统的可用性检测  

Usability Test at System Initialization of HPC

作　　者：寇大治[1] 

机构地区：[1]上海超级计算中心,上海201203

出　　处：《科研信息化技术与应用》2013年第4期76-81,共6页E-science Technology & Application

基　　金：上海市科学技术委员会资助项目(1201H116400);国家高技术研究发展计划(863计划)(2012AA01A308)

摘　　要：本文针对高性能计算机系统在启动/重新启动之后的可用性和健康度的检测方面提出了一套检测标准,这一标准流程自系统启动/重新启动开始,至系统所提供的相关高性能计算应用服务完备为止,包括一系列标准行为以及相应的检测模块。据此设计实现了一套针对曙光5000A超级计算机的检测系统并取得了良好的效果。To ensure that the HPC system is in normal operation and well functioning after system initialization, we proposed a method to monitor the performance of HPC system in this paper. The monitoring system contains several testing modules, the monitoring process started at the beginning of the HPC system initialization, and ends before notifying users that they can login and submit their jobs. Based on the proposed method, we designed and implemented a monitoring system for Dawning 5000A supercomputer, and achieved good testing results.

关 键 词：高性能计算系统 集群系统 可用性检测 

分 类 号：TP338.4[自动化与计算机技术—计算机系统结构]