期刊文献+

PPAT:一种Pthread并行程序线程性能分析工具 被引量:3

PPAT: A THREAD PERFORMANCE ANALYSING TOOL FOR PTHREAD PARALLEL APPLICATIONS
在线阅读 下载PDF
导出
摘要 随着多核/众核处理器技术的快速发展,程序需要越来越多地采用多线程并行技术以提升性能。随着线程个数的增多,线程并行运行过程中相互间同步/互斥及资源竞争关系更加复杂,导致程序性能优化的难度增大。为了使编程人员直观地了解线程的动态运行过程,特别是线程间同步及资源共享带来的影响,帮助其进行程序性能优化,设计实现了一种面向Pthread的并行程序线程性能分析工具PPAT(Pthreads program analysis tool),该工具可在程序运行过程中动态获取线程运行及线程间互斥/同步信息,生成线程通信图,并以多种可视化的方法显示,为编程人员优化程序性能提供依据。 With the development of multi/many-core processors' technology,growing multithreading parallel techniques will be required by the applications in order to enhance performances.However,along with the increase of threads number,the relations of synchronisation and mutual exclusion as well as resource competition in the process of parallel threads executions become more complex,and make the performance tuning more difficult.In this paper we design and implement the PPAT,a parallel program thread performance analysing tool for Pthread applications,in order to have the programmers intuitively learn the dynamic running process of the threads,in particular the influence affected by the synchronisation and resource sharing between the threads,and to assist them in program performance tuning.The tool can obtain dynamically during the program runtime the information of threads execution and mutual exclusion/synchronisation between threads,generate thread communication graph,and display in multiple visualisation approaches,these provide programmers a basis of program performance tuning.
出处 《计算机应用与软件》 CSCD 北大核心 2012年第11期43-47,115,共6页 Computer Applications and Software
基金 国家自然科学基金项目(61073011 61133004) 国家高技术研究发展计划项目(2011AA01A205)
关键词 众核处理器 多线程 资源共享 资源竞争 性能优化 Many-core processor Multithreading Shared data Conflict Performance tuning
  • 相关文献

参考文献16

  • 1Corbalan J, Martorell X, Labarta J. Performance-driven processor allocation [ J ]. IEEE Transactions on Parallel and Distributed Systems ,2005.
  • 2Kumar R,Tullsen D M, Ranganathan P, et al. Single-ISA heterogeneous multi-core architectures for multithreaded workload performance [ J ]. ACM SIGARCH Computer Architecture News,200d.
  • 3Lakshiminarayana N, Rao S, Kim H. Asymmetry aware scheduling algorithms for asymmetric processors. In Workshop on the Interaction between Operating Systems and Computer Architecture[OL]. 2009.
  • 4Suleman M A, Qureshi M, Patt Y. Feedback driven threading: power-effcient and high-performance execution of multithreaded workloads on CMPs[ C ]//16th International Conference on Architectural Support for Programming Languages and Operating Systems ,2008:277 -286.
  • 5Curtis-Maury M, Dzieraa J, Antonopoulos C D, et al. Online power-performance adaptation of muhithreaded programs using hardware event-based prediction [ C ]//Proc. of the 2006 International Conference on Supereomputing,2006.
  • 6Wu P, Kejariwal A, et al. Compiler-driven dependence profiling to guide program paraUelization [ C ]//The 21rd International Workshop on Languages and Compilers for Parallel Computing(LCPC) ,2008.
  • 7Lee J, Wu H, Ravichandran M, et al. Thread tailor: dynamically weaving threads together for efficient, adaptive parallel applications [ J ]. ACM SIGARCH Computer Architecture News,2010.
  • 8Bach M, Charney M, Cohn R, et al. Analyzing parallel programs with PIN [ M ]. IEEE Computer Society ,2010.
  • 9Patil H, Pereira C, Stallcup M, et al. PinPlay : A framework for deterministic replay and reproducible analysis of parallel programs [ C ]// CGO' 10 Proceedings of the 8th annual IEEE/ACM international sym- posium on Code generation and optimization,2010:2-11.
  • 10Adve V S, Vernon M K. Parallel program performance prediction using deterministic task graph analysis [ J ]. ACM Transactions on Computer Systems ,2004:94 - 136.

同被引文献14

引证文献3

二级引证文献7

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