通用图形处理器功耗估算模型

doi:10.3969/j.issn.1000-3428.2017.02.016

摘要/Abstract

摘要： 为精准快速地获得GPU功耗数据,提出一种基于硬件性能计数事件的通用图形处理器(GPGPU)功耗估算方法。通过分析GPGPU程序运行时的功耗分布情况,选择一组与应用程序运行功耗密切相关的硬件性能计数事件集合,使用反向传播人工神经网络分析硬件性能计数事件与实时功耗间的关系,最终建立GPGPU功耗估算模型。实验结果表明,与多元线性回归的功耗估算模型相比,该模型具有更高的估算准确性和通用性。

关键词: 通用图形处理器, 硬件性能计数事件, 反向传播人工神经网络, 交叉验证, 功耗估算

Abstract: In order to get the GPU power data quickly and accurately,this paper proposes a General Purpose Graphics Processing Unit(GPGPU) power estimation model based on hardware performance counting events.Through analysing power distribution during GPGPU program running,it selects a set of performance events which are closely related to application program running power.Then it figures out the relationship between hardware performance counting events and realtime power using Back Propagation Atificial Neural Network(BP-ANN).Finally,it builds a GPGPU power estimation model.Experimental results indicate that compared with the Multiple Linear Regression(MLR) power estimation model,the proposed model has higher estimation accuracy and versatility.

Key words: General Purpose Graphics Processing Unit(GPGPU), hardware performance counting event, Back Propagation Atificial Neural Network(BP-ANN), cross-validation, power estimation

中图分类号:

TP303

王吉军,程华. 通用图形处理器功耗估算模型[J]. 计算机工程, doi: 10.3969/j.issn.1000-3428.2017.02.016.

WANG Jijun,CHENG Hua. Power Estimation Model of General Purpose Graphics Processing Unit[J]. Computer Engineering, doi: 10.3969/j.issn.1000-3428.2017.02.016.

https://www.ecice06.com/CN/Y2017/V43/I2/92

参考文献

参考文献［1］Guan Qingfeng,Shi Xuan,Huang Maoqing,et al.A Hybrid Parallel Cellular Automata Model for Urban Growth Simulation over GPU/CPU Heterogeneous Architectures［J］.International Journal of Geographical Information Science,2016,30(3):494-514. ［2］Ge R,Vogt R,Majumder J,et al.Effects of Dynamic Voltage and Frequency Scaling on a K20 GPU［C］//Proceedings of the 42nd International Conference on Parallel Processing.Washington D.C.,USA:IEEE Computer Society,2013:826-833. ［3］Leng Jingwen,Hetherington T,Eltantawy A,et al.GPUWattch:Enabling Energy Optimizations in GPGPUs［J］.ACM SIGARCH Computer Architecture News,2013,41(3):487-498. ［4］Economou D,Rivoire S,Kozyrakis C,et al.Full-system Power Analysis and Modeling for Server Environ-ments［C］//Proceedings of International Symposium on Computer Architecture.Washington D.C.,USA:IEEE Press,2006:96-100. ［5］Contreras G,Martonosi M.Power Prediction for Intel XScale Processors Using Performance Monitoring Unit Events［C］//Proceedings of International Symposium on Low Power Electronics and Design.Washington D.C.,USA:IEEE Press,2005:221-226. ［6］Joseph R,Martonosi M.Run-time Power Estimation in High Performance Microprocessors［C］//Proceedings of 2001 International Symposium on Low Power Elec-tronics and Design.New York,USA:ACM Press,2001:135-140. ［7］Isci C,Martonosi M.Runtime Power Monitoring in High-end Processors:Methodology and Empirical Data［C］//Proceed-ings of the 36th Annual IEEE/ACM International Symposium on Microarchitecture.Washington D.C.,USA:IEEE Computer Society,2003:93-97. ［8］James D.Nvidia Kepler Architecture［M］.［S.l.］:APC,2012. ［9］Sanders J,Kandrot E.CUDA by Example:An Introduction to General-purpose GPU Programming,Portable Documents［M］.Indianapolis,USA:Addison-Wesley Professional,2010. ［10］Ma Xiaohan,Dong Mian,Zhong Lin,et al.Statistical Power Consumption Analysis and Modeling for GPU-based Computing［C］//Proceedings of ACM SOSP Workshop on Power Aware Computing and Systems.New York,USA:ACM Press,2009:94-98. ［11］Nagasaka H,Maruyama N,Nukada A,et al.Statistical Power Modeling of GPU Kernels Using Performance Counters［C］//Proceeding of 2010 International Green Computing Conference.Washington D.C.,USA:IEEE Press,2010:115-122. ［12］Hong S,Kim H.An Integrated GPU Power and Performance Model［J］.ACM SIGARCH Computer Architecture News,2010,38(3):280-289. ［13］王海峰,陈庆奎.静态程序切片的GPU通用计算功耗预测模型［J］.软件学报,2013,24(8):1746-1760. ［14］王桂彬.基于硬件性能计数器的GPU功耗预测模型［J］.计算机工程与科学,2012,34(3):46-50. ［15］Wittenbrink C M,Kilgariff E,Prabhu A.Fermi GF100 GPU Architecture［J］.IEEE Micro,2011,31(2):50-59. ［16］程华,陈左宁,孙凝晖,等.一种基于细粒度性能计数器建立系统实时功耗模型的方法［J］.计算机学报,2014,37(3):687-699. (下转第104页) (上接第97页) ［17］Curtis-Maury M,Shah A,Blagojevic F,et al.Prediction Models for Multi-dimensional Power-performance Optimiza-tion on Many Cores［C］//Proceedings of the 17th Inter-national Conference on Parallel Architectures and Compila-tion Techniques.New York,USA:ACM Press,2008:250-259. ［18］CUDA Toolkit Documentation［EB/OL］.［2015-09-01］.http://docs.nvidia.com/cuda/cupti/index.html#axzz4 0mOWtBD7. ［19］Eberly L E.Multiple Linear Regression［J］.Topics in Biostatistics,2007,23(2):165-187. ［20］徐勇军,陈静华,骆祖莹,等.CMOS电路动静态功耗协同分析［J］.计算机工程,2006,32(10):231-233. ［21］田景文,高美娟.人工神经网络算法研究及应用［M］.北京:北京理工大学出版社,2006. ［22］Golub G H,Heath M,Wahba G.Generalized Cross-validation as a Method for Choosing a Good Ridge Parameter［J］.Technometrics,1979,21(2):215-223. 编辑陆燕菲

[1]	常婷婷, 翟江涛, 戴跃伟. 一种基于Xgboost的Skype时间式隐信道检测方法[J]. 计算机工程, 2021, 47(7): 88-94.
[2]	王嘉庆, 梅礼晔, 张俊华. 基于深度学习的手骨X射线图像骨龄评估[J]. 计算机工程, 2021, 47(1): 291-297.
[3]	段晓毅, 陈东, 高献伟, 范晓红, 周玉坤. 功耗分析攻击中机器学习模型选择研究[J]. 计算机工程, 2019, 45(11): 144-151,158.
[4]	周文怡, 顾徐波, 施勇. 基于机器学习的网页暗链检测方法[J]. 计算机工程, 2018, 44(10): 22-27.
[5]	邱仁博,娄震. 一种改进的带参数AdaBoost算法[J]. 计算机工程, 2016, 42(7): 199-202,208.
[6]	厉琼莹,夏银水,张骏立. 基于ROBDD的电路功耗估算方法[J]. 计算机工程, 2016, 42(12): 78-83.
[7]	杨柳松, 何光宇. 基于改进粒子群优化的SVM故障诊断方法[J]. 计算机工程, 2013, 39(3): 187-190,196.
[8]	蒋华, 戚玉顺. 基于球结构SVM的多标签分类[J]. 计算机工程, 2013, 39(1): 294-297.
[9]	姚雪琴, 李晓华, 周激流. 基于边缘对称性和HOG的行人检测方法[J]. 计算机工程, 2012, 38(5): 179-182.
[10]	徐晨, 曹辉, 赵晓. 基于SVM的说话人识别参数选择方法[J]. 计算机工程, 2012, 38(21): 175-177.
[11]	刘振华, 傅山. 基于视觉监控的操作手势轨迹分布模式研究[J]. 计算机工程, 2012, 38(15): 142-144,147.
[12]	程克非, 程蕾, 黄永东. 基于J48决策树算法的水质评价方法[J]. 计算机工程, 2012, 38(11): 264-267.
[13]	刘继红, 贾振红, 覃锡忠, 杨杰, 胡英杰. 基于权重粒子群优化阈值的NSCT图像去噪[J]. 计算机工程, 2012, 38(10): 209-211.
[14]	马海晨, 韦刚, 吴百峰. 基于GPGPU的生物序列快速比对[J]. 计算机工程, 2012, 38(04): 241-244.
[15]	蔡娜娜;陈月辉;李　伟. 基于神经网络的蛋白质三级结构预测[J]. 计算机工程, 2010, 36(9): 176-177.

选择文件类型/文献管理软件名称

选择包含的内容