程序能耗测量分析工具FPowerTool及其能耗优化实践

doi:10.3778/j.issn.1673-9418.2102046

计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (6): 1291-1303.DOI: 10.3778/j.issn.1673-9418.2102046

程序能耗测量分析工具FPowerTool及其能耗优化实践

魏光¹, 钱德沛¹^,², 杨海龙¹^,², 栾钟治¹^,²^,⁺()

1. 北京航空航天大学计算机学院中德联合软件研究所,北京 100191
2. 北京航空航天大学软件开发环境国家重点实验室,北京 100191

收稿日期:2021-02-22 修回日期:2021-04-25 出版日期:2022-06-01 发布日期:2021-05-10
通讯作者: + E-mail: luan.zhongzhi@buaa.edu.cn
作者简介:魏光（1986—）,男,博士研究生,主要研究方向为高性能计算、能耗优化等。
钱德沛（1952—）,男,教授,博士生导师,CCF会士,主要研究方向为计算机系统结构、分布式计算、高性能计算、并行计算。
杨海龙（1985—）,男,博士,副教授,博士生导师,CCF高级会员,主要研究方向为高性能计算、性能分析与优化、分布式/并行计算、计算机系统结构。
栾钟治（1971—）,男,博士,副教授,博士生导师,CCF高级会员,主要研究方向为计算机系统结构、高性能计算、分布式计算、并行计算。
基金资助:
国家重点研发计划(2017YFB0202202)

Practice on Program Energy Consumption Optimization by Energy Measurement and Analysis Using FPowerTool

WEI Guang¹, QIAN Depei¹^,², YANG Hailong¹^,², LUAN Zhongzhi¹^,²^,⁺()

1. Sino-German Joint Software Institute, School of Computer Science and Engineering, Beihang University, Beijing 100191, China
2. State Key Laboratory of Software Development Environment, Beihang University, Beijing 100191, China

Received:2021-02-22 Revised:2021-04-25 Online:2022-06-01 Published:2021-05-10
About author:WEI Guang, born in 1986, Ph.D. candidate. His research interests include high-performance com-puting, energy optimization, etc.
QIAN Depei, born in 1952, professor,Ph.D. su-pervisor, fellow of CCF. His research interests include computer architecture, distributed com-puting, high-performance computing and parallel computing.
YANG Hailong, born in 1985, Ph.D., associate professor, Ph.D. supervisor, senior member of CCF. His research interests include high-performance computing, performance analysis and optimiza-tion, distributed/parallel computing and compu-ter architecture.
LUAN Zhongzhi, born in 1971, Ph.D., associate professor, Ph.D. supervisor, senior member of CCF. His research interests include computer architecture, high-performance computing, distri-buted computing and parallel computing.
Supported by:
National Key Research and Development Program of China(2017YFB0202202)

摘要/Abstract

摘要：

能耗感知编程（EAP）是通过优化软件的能效来降低计算系统能耗的一种新途径。它把能耗作为主要指标引入软件开发的过程,通过调整代码的编写方式,降低程序的执行能耗。能耗感知编程的实现面临能耗热点发现、耗能原因确定和问题代码定位等问题。为了解决这些问题,提出了一种程序能耗与性能事件协同测量与分析的新方法EPC。首先,简要论述EPC的基本原理和基于该方法的程序能耗测量与分析工具FPowerTool的实现。然后,介绍如何对能耗和性能事件做关联分析,找出影响程序能耗的主要因素。最后,通过一组程序优化案例,说明如何通过关联分析定位与高能耗相关的程序代码,进而改变代码编写和数据放置与访问方式,使程序的执行能耗得到降低。实验结果表明,基于EPC方法提供的能耗感知和分析能力,通过改进数据定义、赋值、放置与访问方式等可以改善程序性能和能效。

关键词: 能耗感知, 能耗优化, 性能事件, 能耗-性能关联

Abstract:

Energy-aware programming (EAP) is a new approach to reduce energy consumption of computing systems. It introduces energy as one of the main design metrics into the process of software development to reduce program energy consumption by adjusting the way of programming. The implementation of EAP is facing some difficulties in finding energy consumption hot spots, identifying main factors which cause excessive energy consumption, and locating inappropriate code segments in the program. To address these issues, this paper proposes a new method called EPC (energy-performance correlation) for joint measurement and analysis of energy consumption and perfor-mance events during program execution. Firstly, the basic principles of EPC are introduced and the implementation of an EPC-based tool, FPowerTool, for program energy consumption measurement and analysis is presented. Then, the method of energy-performance events correlation analysis for identifying the main factors influencing energy consumption is presented. Finally, a set of programs is used as case studies to show how to locate the code segments related to high energy consumption by correlation analysis, and how to change the coding and data placement and access to reduce the program energy consumption. The experiment results show that based on the energy-aware and analysis capabilities provided by the EPC method, program performance and energy efficiency can be improved by improving data definition, assignment, placement, and access methods.

Key words: energy-aware, energy consumption optimization, performance event, energy-performance correlation

中图分类号:

TP311

魏光, 钱德沛, 杨海龙, 栾钟治. 程序能耗测量分析工具FPowerTool及其能耗优化实践[J]. 计算机科学与探索, 2022, 16(6): 1291-1303.

WEI Guang, QIAN Depei, YANG Hailong, LUAN Zhongzhi. Practice on Program Energy Consumption Optimization by Energy Measurement and Analysis Using FPowerTool[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(6): 1291-1303.

图/表 20

参考文献 34

[1]	WEI G, QIAN D, YANG H, et al. FPowerTool: a function-level power profiling tool[J]. IEEE Access, 2019, 7: 185710-185719. DOI URL
[2]	DAVID H, GORBATOV E, HANEBUTTE U R, et al. RAPL: memory power estimation and capping[C]// Proceedings of the 2010 ACM/IEEE International Symposium on Low-Power Electronics and Design, Austin, Aug 18-20, 2010. New York: ACM, 2010: 189-194.
[3]	Intel. Intel^® 64 and IA-32 architectures software developer’s manual, volume 3b: system programming guide[R]. 2016: 33.
[4]	WEAVER V M, JOHNSON M, KASICHAYANULA K, et al. Measuring energy and power with PAPI[C]// Proceedings of the 41st International Conference on Parallel Processing Workshops, Pittsburgh, Sep 10-13, 2012. Washington: IEEE Computer Society, 2012: 262-268.
[5]	MCCRAW H, RALPH J, DANALIS A, et al. Power moni-toring with PAPI for extreme scale architectures and dataflow-based programming models[C]// Proceedings of the 2014 IEEE International Conference on Cluster Computing, Ma-drid, Sep 22-26, 2014. Washington: IEEE Computer Society, 2014: 385-391.
[6]	KUFRIN R. PerfSuite: an accessible, open source, performance analysis environment for Linux[C]// Proceedings of the 6th International Conference on Linux Clusters, Chapel Hill, 2005.
[7]	KUFRIN R. Measuring and improving application performance with perfsuite[J]. Linux Journal, 2005(135): 4.
[8]	ENBODY R. Perfmon: performance monitoring tool[D]. Lan-sing: Michigan State University, 1999.
[9]	ERANIAN S. Perfmon: Linux performance monitoring for IA-64[EB/OL]. [2020-12-14]. http://www.hpl.hp.com/research/linux/perfmon .
[10]	ERANIAN S. Perfmon2: a flexible performance monitoring interface for Linux[C]// Proceedings of the 2006 Ottawa Linux Symposium, Ottawa, Jul 2006: 269-288.
[11]	DE MELO A C. The new Linux “perf” tools[R]. Brazil: Red Hat Inc., 2010.
[12]	TREIBIG J, HAGER G, WELLEIN G. LIKWID:a light-weight performance-oriented tool suite for x86 multicore environments[C]// Proceedings of the 39th International Con-ference on Parallel Processing, San Diego, Sep 13-16, 2010. Washington: IEEE Computer Society, 2010: 207-216.
[13]	EULISSE G, TUURA L A. IgProf profiling tool[R]. Bosto: Northeastern University, 2005: 1-4.
[14]	KHAN K N, NYBACK F, OU Z H, et al. Energy profiling using IgProf[C]// Proceedings of the 15th IEEE/ACM Inter-national Symposium on Cluster, Cloud and Grid Compu-ting, Shenzhen, May 4-7, 2015. Washington: IEEE Computer Society, 2015: 1115-1118.
[15]	MANOUSAKIS I, ZAKKAK F S, PRATIKAKIS P, et al. TProf: an energy profiler for task-parallel programs[J]. Sus-tainable Computing: Informatics and Systems, 2015, 5: 1-13.
[16]	MUKHANOV L, PARASYRIS N, WANG Z, et al. ALEA: a fine-grained energy profiling tool[J]. ACM Transactions on Architecture and Code Optimization, 2017, 14(1): 1-25.
[17]	MUKHANOV L, NIKOLOPOULOS D S, DE SUPINSKI B R. ALEA: fine-grain energy profiling with basic block sampling[C]// Proceedings of the 2015 International Confe-rence on Parallel Architectures and Compilation, San Fran-cisco, Oct 18-21, 2015. Washington: IEEE Computer Society, 2015: 87-98.
[18]	SMEJKAL T, HÄHNEL M, ILSCHE T, et al. E-Team: prac-tical energy accounting for multi-core systems[C]// Procee-dings of the 2017 USENIX Annual Technical Conference, Santa Clara, Jul 12-14, 2017: 589-601.
[19]	宋杰, 孙宗哲, 李甜甜, 等. 面向代码的软件能耗优化研究进展[J]. 计算机学报, 2016, 39(11): 2270-2290.
	SONG J, SUN Z Z, LI T T, et al. Research advance on code oriented optimization of software energy consumption[J]. Chinese Journal of Computers, 2016, 39(11): 2270-2290.
[20]	KANDEMIR M T, KOLCU I, KADAYIF I. Influence of loop optimizations on energy consumption of multi-bank memory systems[C]// LNCS 2304: Proceedings of the 11th International Conference on Compiler Construction, Grenoble, Apr 8-12, 2002. Berlin, Heidelberg: Springer, 2002: 276-292.
[21]	BUNSE C, HÖPFNER H, MANSOUR E, et al. Exploring the energy consumption of data sorting algorithms in em-bedded and mobile environments[C]// Proceedings of the 2009 10th International Conference on Mobile Data Mana-gement: Systems, Services and Middleware, Taipei, China, May 18-20, 2009. Washington: IEEE Computer Society, 2009: 600-607.
[22]	DAYARATHNA M, WEN Y, MEMBER S, et al. Data center energy consumption modeling: a survey[J]. IEEE Com-munications Surveys & Tutorials, 2016, 18(1): 732-794.
[23]	O’BRIEN K, PIETRI I, REDDY R, et al. A survey of power and energy predictive models in HPC systems and appli-cations[J]. ACM Computing Surveys, 2017, 50(3): 37.
[24]	ZANG W, GORDON-ROSS A. A survey on cache tuning from a power/energy perspective[J]. ACM Computing Sur-veys, 2013, 45(3): 32.
[25]	MONCHIERO M, CANAL R, GONZÁLEZ A. Design space exploration for multicore architectures: a power/performance/thermal view[C]// Proceedings of the 20th Annual Interna-tional Conference on Supercomputing, Cairns, Jun 28-Jul 1, 2006. New York: ACM, 2006: 177-186.
[26]	KOWARSCHIK M, WEISS C. An overview of cache opti-mization techniques and cache-aware numerical algorithms[C]// LNCS 2625: Proceedings of the Algorithms for Memory Hierarchies, Mar 10-14, 2002. Berlin, Heidelberg: Springer, 2003: 213-232.
[27]	WEI G, QIAN D P, YANG H L, et al. Modeling power con-sumption of the code execution using performance counters statistics[C]// Proceedings of the 20th International Confe-rence on Parallel and Distributed Computing, Applications and Technologies, Gold Coast, Dec 5-7, 2019. Piscataway: IEEE, 2019: 381-385.
[28]	BIENIA C, KUMAR S, SINGH J P, et al. The PARSEC ben-chmark suite: characterization and architectural implications[C]// Proceedings of the 17th International Conference on Parallel Architectures and Compilation Techniques, Toronto, Oct 25-29, 2008. New York: ACM, 2008: 72-81.
[29]	BAILEY D H, BARSZCZ E, BARTON J T, et al. The NAS parallel benchmarks[J]. The International Journal of Super-computing Applications, 1991, 5(3): 63-73.
[30]	IZENMAN A J. Linear discriminant analysis[M]// Modern Multivariate Statistical Techniques. Berlin, Heidelberg: Sp-ringer, 2013: 237-280.
[31]	CHE S, BOYER M, MENG J Y, et al. Rodinia: a bench-mark suite for heterogeneous computing[C]// Proceedings of the 2009 IEEE International Symposium on Workload Cha-racterization, Austin, Oct 4-6, 2009. Washington: IEEE Com-puter Society, 2009: 44-54.
[32]	Rogue Wave Software. CPU cache optimization: does it matter? Should I worry? Why? An exploration of the world of CPU cache performance[R/OL]. [2020-12-14]. https://studylib.net/doc/18141740/cpu-cache-optimization--does-it-matter .
[33]	LAROS III J H, PEDRETTI K T, KELLY S M, et al. Energy-efficient high performance computing: measurement and tuning[M]. Berlin, Heidelberg: Springer, 2012.
[34]	HOROWITZ M, INDERMAUR T, GONZALEZ R. Low-power digital design[C]// Proceedings of the 1994 IEEE Symposium on Low Power Electronics, San Diego, Oct 10-12, 1994. Piscataway: IEEE, 1994: 8-11.

类别	模型用到的性能事件
硬件事件	buscycles, cachereferences, cpucycles, instructions, refcycles
软件事件	cpuclock, cpumigrations, minorfaults, taskclock
硬件cache 事件	L1dcacheloadmisses, L1dcachestores, L1icacheloadmisses, LLCloads, branchloads, dTLBloadmisses, dTLBstoremisses, iTLBloads, nodeloadmisses, nodeloads, nodestores

类别	模型用到的性能事件
硬件事件	buscycles, cachereferences, cpucycles, instructions, refcycles
软件事件	cpuclock, cpumigrations, minorfaults, taskclock
硬件cache 事件	L1dcacheloadmisses, L1dcachestores, L1icacheloadmisses, LLCloads, branchloads, dTLBloadmisses, dTLBstoremisses, iTLBloads, nodeloadmisses, nodeloads, nodestores

性能事件	相对重要性	性能事件	相对重要性
cpucycles	4.386	nodeloads	0.223
cpuclock	4.326	dTLBloadmisses	0.220
taskclock	3.855	dTLBstoremisses	0.216
refcycles	2.400	L1dcacheloadmisses	0.187
buscycles	2.161	L1dcachestores	0.151
minorfaults	0.864	nodeloadmisses	0.097
iTLBloads	0.773	cachereferences	0.090
instructions	0.715	nodestores	0.077
L1icacheloadmisses	0.481	cpumigrations	0.051
LLCloads	0.423	branchloads	0.019

性能事件	相对重要性	性能事件	相对重要性
cpucycles	4.386	nodeloads	0.223
cpuclock	4.326	dTLBloadmisses	0.220
taskclock	3.855	dTLBstoremisses	0.216
refcycles	2.400	L1dcacheloadmisses	0.187
buscycles	2.161	L1dcachestores	0.151
minorfaults	0.864	nodeloadmisses	0.097
iTLBloads	0.773	cachereferences	0.090
instructions	0.715	nodestores	0.077
L1icacheloadmisses	0.481	cpumigrations	0.051
LLCloads	0.423	branchloads	0.019

服务器	配置
CPU	Intel^® Xeon^® CPU E5-2680 v3 2 sockets, 12 cores per socket, 2.50 GHz
操作系统	CentOS 7.3.1611
内存	125 GB
编译器	gcc 4.8.5
Cache信息	L1d cache: 32 KB L1i cache: 32 KB L2 cache: 256 KB L3 cache: 30 720 KB

程序能耗测量分析工具FPowerTool及其能耗优化实践

Practice on Program Energy Consumption Optimization by Energy Measurement and Analysis Using FPowerTool

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 20

参考文献 34

相关文章 15

编辑推荐

Metrics

类别	原生程序	优化后	差值	差值百分比/%
时间/μs		5.58E+06	4.91E+06	-6.70E+05	-12.0
能量/J	PACKAGE0	175.2	158.2	-17.0	-9.7
	PACKAGE1	207.7	188.7	-19.0	-9.1
	DRAM0	17.4	14.9	-2.5	-14.4
	DRAM1	34.2	31.7	-2.5	-7.3

性能事件	原生程序	优化后	差值	差值百分比/%
branch_instructions	4.50E+09	4.23E+09	-2.70E+08	-6.0
branch_misses	4.29E+07	4.35E+07	6.00E+05	1.4
bus_cycles	6.12E+08	5.64E+08	-4.80E+07	-7.8
cache_misses	2.83E+04	2.11E+04	-7.20E+03	-25.4
cache_references	8.84E+04	8.61E+04	-2.30E+03	-2.6
cpu_cycles	1.77E+10	1.61E+10	-1.60E+09	-9.0
instructions	4.11E+10	3.80E+10	-3.10E+09	-7.5
ref_cpu_cycles	1.53E+10	1.41E+10	-1.20E+09	-7.8
cpu_clock	6.14E+09	5.65E+09	-4.90E+08	-8.0
page_faults_min	5.03E+02	4.76E+02	-2.70E+01	-5.4
task_clock	6.14E+09	5.65E+09	-4.90E+08	-8.0

类别	程序1	程序2
时间/μs		7.60E+04	3.67E+04
能量/J	PACKAGE0	1.54	0.70
	PACKAGE1	1.63	0.83
	DRAM0	0.22	0.10
	DRAM1	0.42	0.21

硬件事件和软件事件	程序1	程序2	硬件cache事件	程序1	程序2
branch_instructions	1.14E+07	1.11E+07	dtlbreadaccess	2.23E+06	2.23E+06
branch_misses	4.38E+03	1.50E+03	dtlbreadmiss	1.38E+02	2.06E+02
bus_cycles	7.60E+06	4.37E+06	dtlbwriteaccess	5.48E+06	4.49E+06
cache_misses	9.93E+04	5.03E+04	dtlbwritemiss	3.41E+02	9.80E+01
cache_references	1.47E+06	4.09E+05	itlbreadaccess	4.15E+02	9.90E+01
cpu_cycles	1.17E+08	6.02E+07	itlbreadmiss	1.49E+02	7.40E+01
instructions	5.68E+07	5.54E+07	L1dreadaccess	6.64E+06	5.72E+06
ref_cpu_cycles	1.90E+08	1.09E+08	L1dreadmiss	3.15E+06	1.47E+06
cpu_clock	7.63E+07	4.39E+07	L1dwriteaccess	1.00E+07	7.40E+06
page_faults_min	1.21E+03	1.10E+03	L1ireadmiss	1.95E+04	9.15E+03
task_clock	7.63E+07	4.39E+07	LLreadaccess	8.19E+05	2.09E+05
			LLreadmiss	8.85E+03	4.59E+03
			LLwriteaccess	1.14E+04	6.43E+03
			LLwritemiss	1.11E+04	5.80E+03

类别	程序3	程序4
时间/μs	1.76E+05	1.30E+05
能量/J	PACKAGE0	3.58	2.75
	PACKAGE1	4.17	2.88
	DRAM0	0.50	0.42
	DRAM1	1.07	0.77

硬件事件和软件事件	程序3	程序4	硬件cache事件	程序3	程序4
branch_instructions	3.95E+07	3.90E+07	dtlbreadaccess	9.00E+06	7.95E+06
branch_misses	3.00E+03	2.36E+03	dtlbreadmiss	2.39E+03	7.28E+02
bus_cycles	1.71E+07	1.31E+07	dtlbwriteaccess	1.98E+07	1.68E+07
cache_misses	2.54E+05	1.70E+05	dtlbwritemiss	1.23E+03	3.01E+02
cache_references	2.85E+06	1.31E+06	itlbreadaccess	9.00E+00	8.00E+00
cpu_cycles	3.32E+08	2.33E+08	itlbreadmiss	4.10E+01	1.70E+01
instructions	1.59E+08	1.57E+08	L1dreadaccess	2.31E+07	2.26E+07
ref_cpu_cycles	4.27E+08	3.27E+08	L1dreadmiss	8.05E+06	5.61E+06
cpu_clock	1.71E+08	1.31E+08	L1dwriteaccess	3.04E+07	2.87E+07
page_faults_min	7.28E+02	6.56E+02	L1ireadmiss	2.75E+04	2.40E+04
task_clock	1.71E+08	1.31E+08	LLreadaccess	1.45E+06	6.49E+05
			LLreadmiss	1.35E+04	9.83E+03
			LLwriteaccess	3.38E+04	2.07E+04
			LLwritemiss	3.39E+04	1.99E+04

类别	程序5	程序6
时间/μs	6.85E+04	1.97E+04
能量/J	PACKAGE0	1.49	0.42
	PACKAGE1	1.49	0.41
	DRAM0	0.24	0.08
	DRAM1	0.31	0.11

硬件事件和软件事件	程序5	程序6	硬件cache事件	程序5	程序6
branch_instructions	3.09E+06	3.07E+06	dtlbreadaccess	6.77E+05	7.93E+05
branch_misses	4.62E+03	3.00E+03	dtlbreadmiss	1.26E+03	5.40E+01
bus_cycles	6.86E+06	2.19E+06	dtlbwriteaccess	1.52E+06	1.95E+06
cache_misses	4.22E+05	2.74E+04	dtlbwritemiss	5.14E+02	3.11E+02
cache_references	5.19E+06	2.69E+05	itlbreadaccess	6.56E+02	1.25E+03
cpu_cycles	1.03E+08	2.79E+07	itlbreadmiss	2.20E+01	6.60E+01
instructions	1.82E+07	2.07E+07	L1dreadaccess	1.85E+06	2.12E+06
ref_cpu_cycles	1.72E+08	5.49E+07	L1dreadmiss	4.44E+06	7.54E+05
cpu_clock	6.89E+07	2.20E+07	L1dwriteaccess	2.48E+06	2.43E+06
page_faults_min	1.06E+03	1.06E+03	L1ireadmiss	5.31E+03	4.59E+03
task_clock	6.89E+07	2.20E+07	LLreadaccess	1.54E+06	1.07E+05
			LLreadmiss	2.44E+05	2.19E+03
			LLwriteaccess	3.53E+05	4.17E+03
			LLwritemiss	2.73E+03	4.00E+03

[1]	赵学武, 王红梅, 刘超慧, 李玲玲, 薄树奎, 冀俊忠. 面向人脑功能划分的人工水母搜索优化算法[J]. 计算机科学与探索, 2022, 16(8): 1829-1841.
[2]	郭金鑫, 张广婷, 张云泉, 陈泽华, 贾海鹏. Cooley-Tukey FFT算法高性能实现与优化研究[J]. 计算机科学与探索, 2022, 16(6): 1304-1315.
[3]	毛伊敏, 耿俊豪. 结合信息论和范数的并行随机森林算法[J]. 计算机科学与探索, 2022, 16(5): 1064-1075.
[4]	李金红, 王丽珍, 周丽华. 模糊特征的top-k平均效用co-location模式挖掘[J]. 计算机科学与探索, 2022, 16(5): 1053-1063.
[5]	陈李越, 柴迪, 王乐业. UCTB：时空人群流动预测工具箱[J]. 计算机科学与探索, 2022, 16(4): 835-843.
[6]	张少伟, 王鑫, 陈子睿, 王林, 徐大为, 贾勇哲. 有监督实体关系联合抽取方法研究综述[J]. 计算机科学与探索, 2022, 16(4): 713-733.
[7]	蒋祎莹, 张丽平, 金飞虎, 郝晓红. 空间数据库中混合数据组最近邻查询[J]. 计算机科学与探索, 2022, 16(2): 348-358.
[8]	李朝阳, 李琳, 陶晓辉. 面向动态交通流预测的双流图卷积网络[J]. 计算机科学与探索, 2022, 16(2): 384-394.
[9]	赵恒泰, 赵宇海, 袁野, 季航旭, 乔百友, 王国仁. 分布式环境下大规模维表关联技术优化[J]. 计算机科学与探索, 2022, 16(2): 337-347.
[10]	刘卫明, 张弛, 毛伊敏. 采用N-list结构的混合并行频繁项集挖掘算法[J]. 计算机科学与探索, 2022, 16(1): 120-136.
[11]	宗枫博, 赵宇海, 王国仁, 季航旭. 面向多表数据连接投影和连接顺序的优化方法[J]. 计算机科学与探索, 2022, 16(1): 106-119.
[12]	林洪武 1,2 , 尤朝 1,2 , 周明辉1,2+ , 梅宏 1,2 . 以代理为中心的 OSGi 构件资源监控方法[J]. 计算机科学与探索, 2011, 5(1): 23-31.
[13]	朱小虎1,2 , 宋文军1,2 , 王崇骏1,2+ , 谢俊元1,2 . 用于社团发现的Girvan-Newman改进算法[J]. 计算机科学与探索, 2010, 4(12): 1101-1108.
[14]	曾红卫+, 缪淮扣 . 模型检验在构件数据流测试中的应用[J]. 计算机科学与探索, 2010, 4(12): 1121-1130.
[15]	袁崇义1,2 , 黄雨1,2,3+ , 赵文1,2,3 , 黄舒志2 . 操作表达式的Petri网表示*[J]. 计算机科学与探索, 2010, 4(11): 961-976.