期刊文献+

一个基于硬件计数器的程序性能测试与分析工具 被引量:3

A Hardware Counter Based Tool for Application's Performance Measurement and Analysis
下载PDF
导出
摘要 在IntelP6系列处理器与Microsoft Windows NT平台上开发了一个工具软件PTracker,它利用处理器中的硬件性能计数器来获取程序性能数据,并结合机器体系结构参数对数据进行分析。它无需用户编程,与应用程序所使用的编程语言无关,使用很方便。它不仅能够通过性能计数器获得精确的性能参数,而且还能通过对测试得到的性能数据的分析,揭示程序高层次的性能特征,对程序性能评价与优化具有一定的指导作用。本文介绍了PTracker的技术背景、设计与系统实现,并给出了一个应用实例。 This article introduces PTracker, a useful tool designed and implemented for Intel P6 processors running Microsoft Windows NT operating system. It utilizes hardware performance monitoring counters in modern processors to measure an application's performance data. Furthermore, it performs additional calculation to the measured data based on the target platform's architecture parameters and profiles. PTracker is convenient to use in that it requires no modification to the applications' source codes and is independent of the application's programming language. It not only captures the application's signatures, but also provides more informative performance characteristics of the application, which are helpful for the understanding and tuning of the application. A case of its application in program optimization is presented.
出处 《计算机科学》 CSCD 北大核心 2004年第1期170-174,共5页 Computer Science
基金 国家自然科学基金(69933030)
关键词 硬件计数器 处理器 程序性能 性能测试 数据分析 Hardware performance monitoring counters .Performance measurement .Data analysis .Performance tuning
  • 相关文献

参考文献16

  • 1[1]Ghosh S,et al.Cache Miss Equations: A Compiler Framework for Analyzing and Tuning Memory Behavior.In ACM Transactions on Programming Languages and Systems,1999,21(4):702~745
  • 2[2]http://www.cs.wisc.edu/~mscalar/simplescalar.html
  • 3[3]Merten M C,et al.An Architectural Framework for Run-Time Optimization.IEEE Transactions on Computers,2001,50(6):567~589
  • 4[4]Lambert, et al.Profiling I/O Interrupts in Modern Architectures.In:8th Intl.Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems,San Francisco, California,2000
  • 5[5]Hirzel M, et al.Bursty Tracing: A Framework for Low-Overhead Temporal Profiling.In:4th Workshop on Feedback-Directed and Dynamic Optimization (FDDO), Dec.2001
  • 6[6]http://icl.cs.utk.edu/projects/papi/
  • 7[7]http://www.gz-juelich.de/zam/PCL/
  • 8[8]http://research.compaq.com/SRC/dcpi/
  • 9[9]http://developer.intel.com/vtune/
  • 10[10]IA-32 Intelⅳ Architecture Software Developer's Manual: Volume 3: System Programming Guide.Intel Corporation.2002

同被引文献36

引证文献3

二级引证文献2

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