• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 论文 • 上一篇    下一篇

软件交互性能异常现象分析

何嘉权1,陈渝2   

  1. (1.清华大学软件学院,北京100084;2.清华大学计算机科学与技术系,北京 100084)
  • 收稿日期:2015-04-27 修回日期:2015-09-29 出版日期:2017-09-25 发布日期:2017-09-25
  • 基金资助:

    核高基重大专项(2012ZX01039-004,2013ZX01039001-002)

Software interactive performance anomaly analysis

HE Jia-quan1,CHEN Yu2   

  1. (1.School of Software,Tsinghua University,Beijing 100084;
    2.Department of Computer Science and Technology,Tsinghua University,Beijing 100084,China)

     
  • Received:2015-04-27 Revised:2015-09-29 Online:2017-09-25 Published:2017-09-25

摘要:

近年来,得益于个人计算机乃至移动设备的普及,以及现代操作系统的发展,应用软件开发被推向一个前所未有的热潮。为了提高应用软件的用户体验,多线程程序开发技术已经广泛应用于软件开发的各个环节。多线程程序开发技术,一方面让计算机硬件资源得以充分利用,提高了软件的响应速度,另一方面也增加了程序开发的难度,以及增加了应用软件出现性能异常之后的分析调试难度。考察了现有软件性能异常现象分析的工作,通过对现有的交互性能问题进行扩展,以及对分析模型提出改进,以满足Linux操作系统上的应用软件性能异常场景的分析需要,并在Linux平台上设计交互性能异常分析系统,以Google Chrome和GNOME Nautilus等流行软件为分析对象,分析实际场景中的性能异常现象。
 

关键词: 交互性能异常, 线程依赖, 动态分析

Abstract:

The development of personal computers and mobile devices, as well as modern operating systems, has promoted application software developing to a new level. Multithreading program development is then introduced to optimize software user experience. While it makes full use of hardware and shortens software response latency, it complicates software development and makes it more difficult to analyze and debug. We look into the existing work regarding interactive performance anomaly analysis, broaden the range of previous known interactive performance anomalies, and improve the existing analyzing model to meet the performance anomaly analysis needs of application software in Linux operating system. Based on the mentioned work, we design an interactive performance anomaly analyzing system on Linux, and using popular software such as Google Chrome and GNOME Nautilus as the objectives we analyze some interactive performance anomalies in practice.

Key words: interactive performance anomaly, thread dependency, dynamic analysis