期刊文献+

并行程序性能和功耗的测试分析工具研究

A Survey of Performance and Power Measurement and Analysis Tools for Parallel Programs
下载PDF
导出
摘要 高性能并行计算在社会的各个方面发挥了重要作用,然而计算性能发展所带来的高功耗已经成为阻碍其进一步提高的重要因素。除了在硬件制程工艺等方向进行优化外,从软件角度来提高系统的能量有效性是当前高性能计算领域研究的热点问题。其中对并行程序的性能和功耗特征进行检测分析,可以有助于研究人员了解程序行为,找出程序性能瓶颈和功耗的热点区域,发现程序的能量有效性优化的方向。方便准确的并行程序性能和功耗的测试分析工具对于并行程序开发和能量有效性的优化有着极其重要的作用。该文对并行程序性能和功耗的测试工具进行了对比研究。首先对并行程序性能和功耗测试的原理进行阐述;其次,对当前国外超算中心常用的分析和测量软件进行介绍和对比;最后进行了总结。 High-performance parallel computing has played a major role in all aspects of society,but the high power consumption brought about by the development of computing performance has become an important factor hindering its further improvement.In addition to optimization in the direction of hardware process technology,it is a hot issue in the field of high-performance computing to improve the energy efficiency of the system from the perspective of software.Among them,the measurement and analysis of the performance and power consumption characteristics of parallel programs can help researchers understand program behavior,find out program performance bottlenecks and power consumption hot spots,and find the direction of energy efficiency optimization of programs.Convenient and accurate parallel program performance and power measurement and analysis tools play an extremely influential role in parallel program development and energy efficiency optimization.We summarize the measurement and analysis tools of for performance and power measurement parallel program.Firstly,the principle of performance and power consumption measurement of parallel program is described.Secondly,the software commonly used in foreign supercomputing centers is introduced and compared.Finally,the summary is made.
作者 潘晓东 孙晓乐 郑文旭 吴丹 PAN Xiao-dong;SUN Xiao-le;ZHENG Wen-xu;WU Dan(School of Computer,National University of Defense Technology,Changsha 410073,China)
出处 《计算机技术与发展》 2021年第7期69-74,共6页 Computer Technology and Development
基金 国家重点研发计划(2018YFB1003203)。
关键词 高性能计算 性能分析工具 功耗检测 并行计算 能量有效性 超算中心 high performance computing performance analysis tools power consumption measurement parallel computing energy efficiency supercomputer center
  • 相关文献

参考文献2

二级参考文献17

  • 1张骏,樊晓桠,刘松鹤.多核、多线程处理器的低功耗设计技术研究[J].计算机科学,2007,34(10):301-305. 被引量:15
  • 2CDC 6600.[EB/OL]. http://en.wikipedia.org/wiki/ CDC 6600.
  • 3TOP500.[EB/OL]. [ 2013-11-18 ] .http://www.top500.org/ statistics/list/.
  • 41993年至2013年20年间国际TOP500发布的关于高性能计算应用领域分布的统计数据分析图表.[EB/OL][2013-11-18].http://www.top500.org/statistics/overtime/.
  • 52013年中国高性能计算机性能TOP100排行榜.[EB/OL].[2013-10-29].http://hpcchina2013.csp.escience.cn/dct/page/1.
  • 6多个超算中心联合呼吁:打破中国超算应用瓶颈.[EB/OL].[2012-04-11].http://news.xinhuanet.com/tech/2012-04/11/c_122963148.htm.
  • 7美国橡树岭国家实验室领导计算设施2012-2013年报.[EB/OL].https://www.olcf.ornl.gov/media-center/center-reports/.
  • 8美国劳伦斯利弗莫尔国家实验室超级计算中心.[EB/OL].https://www.llnl.gov/about/mgtsponsors.html.
  • 9日本理化研究所先进科学计算机构.[EB/OL].http://www.riken.jp/en/research/labs/aics/.
  • 10德国于利希超级计算中心.[EB/OL].http://www.fz-juelich.de/portal/EN/AboutUs/FactsFigures/infrastructure/_node.html.

共引文献14

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