摘要
由于并行应用程序的运行效率往往很低 ,如何帮助程序员提高性能就成为高性能计算中的重要问题 .本文介绍了一个基于 MPI的性能评价工具 ,它可以在应用程序运行的同时收集系统负载信息 ,跟踪程序流程 ,根据硬件资源情况对处理机进行分组 ,并将负载信息和程序流程同时以图形方式展示 .程序员可以藉此对并行应用程序运行情况进行监测 ,分析算法执行过程和系统负载的关系 ,找出性能瓶颈 ,发掘应用程序的潜力 ,最终提高应用程序的性能 .
Improving the low efficiency of most parallel applications is an important issue in high performance computing. This paper presents a Performance Tool(PT) on the base of MPI, which can collect the information of CPU load and memory utilization during runtime, trace the process of application, divide computers into groups according to their hardware resource, and visualize the information collected in screen. With this performance tool, programmers can monitor the runtime state of the application, analyze the relationship between program process and system load, find out the performance bottleneck, and improve the performance of the application at last.
出处
《小型微型计算机系统》
CSCD
北大核心
2002年第8期902-906,共5页
Journal of Chinese Computer Systems
