计算机科学 ›› 2022, Vol. 49 ›› Issue (10): 59-65.doi: 10.11896/jsjkx.210800103

• 高性能计算* 上一篇    下一篇

ARM架构云服务器的CPU功耗模型研究

金育妍1, 余天豪1, 王松波1, 林伟伟1,3, 潘宇聪2   

  1. 1 华南理工大学计算机科学与工程学院 广州 510641
    2 广东省高性能计算重点实验室 广州 510033
    3 鹏程实验室 广东 深圳 518066
  • 收稿日期:2021-08-12 修回日期:2022-03-02 出版日期:2022-10-15 发布日期:2022-10-13
  • 通讯作者: 林伟伟(linww@scut.edu.cn)
  • 作者简介:(jingyy.kk@outlook.com)
  • 基金资助:
    国家自然科学基金(62072187);广东省重点研发计划(2021B0101420002);广东省基础与应用基础研究重大项目(2019B030302002);广州市科技计划(202007040002);广州市开发区国际合作项目(2020GH10);广东省高性能计算重点实验室2019年课题资助;鹏程实验室重大任务项目(PCL2012A09)

CPU Power Model for ARM Architecture Cloud Servers

JIN Yu-yan1, YU Tian-hao1, WANG Song-bo1, LIN Wei-wei1,3, PAN Yu-cong2   

  1. 1 School of Computer Science and Engineering,South China University of Technology,Guangzhou 510641,China
    2 Guangdong Science and Technology Infrastructure Center,Guangzhou 510033,China
    3 Pengcheng Laboratory,Shenzhen,Guangdong 518066,China
  • Received:2021-08-12 Revised:2022-03-02 Online:2022-10-15 Published:2022-10-13
  • About author:JIN Yu-yan,born in 1998,postgra-duate.Her main research interests include cloud computing,big data technology and resource scheduling.
    LIN Wei-wei,born in 1980,Ph.D,professor,is a senior member of China Computer Federation.His main research interests include cloud computing,big data technology and AI application technology.
  • Supported by:
    National Natural Science Foundation of China(62072187),Key-Area Research and Development Program of Guangdong Province(2021B0101420002),Guangdong Major Project of Basic and Applied Basic Research(2019B030302002),Guangzhou Science and Technology Plan(202007040002),Guangzhou Development Zone International Cooperation Project(2020GH10),Guangdong Provincial Key Laboratory of High Performance Computing and 2019 Project Funding and Major Key Project of PCL(PCL2012A09).

摘要: 云服务器的功耗模型是云数据中心能耗优化研究的重要内容之一。CPU功耗模型是云服务器功耗模型的重要组成部分,然而现有CPU功耗模型没有考虑CPU的异构性,如缺乏对ARM架构服务器CPU功耗模型的研究。在调研分析现有的ARM架构CPU功耗模型的基础上,提出了一种面向ARM架构的新CPU功耗模型——基于混合建模的CPU功耗模型(Hybrid Based Model,HBM)。该功耗模型综合考虑了CPU利用率和CPU性能事件等建模特征,相比现有的测算精度很高的基于性能计数器的CPU功耗模型,HBM的测算精度与其相近且模型训练成本更低,更适合ARM服务器的CPU功耗建模。文中使用Sysbench负载工具对所提HBM进行实验验证,实验结果表明,HBM的平均相对误差(MRE)在1%以内,具有良好的测算精度。此外,还针对x86和ARM架构服务器进行了交叉实验,实验结果表明不同架构服务器的CPU功耗行为相异,应当使用不同的CPU功耗建模方法。

关键词: 能耗优化, 功耗模型, CPU异构性, ARM架构

Abstract: The power model of cloud server is one of the important contents of the research on the energy consumption optimization of cloud data center.The CPU power model is an important part of the power models of cloud servers.However,the existing CPU power models do not consider the CPU heterogeneity,such as lack of research on the CPU power model of ARM architecture cloud servers.Based on the investigation and analysis of existing ARM architecture CPU power models,this paper proposes a new CPU power model oriented to the ARM architecture,namely the hybrid based model(HBM).HBM comprehensively considers modeling features such as CPU utilization and CPU performance events.Compared with existing PMC based model with high measurement accuracy,HBM has similar measurement accuracy and lower model training cost.Thus,HBM is more suitable for CPU power modeling of ARM servers.This paper uses the Sysbench benchmark to verify HBM,and experimental results show that the mean relative error(MRE) of HBM is within 1%,which means HBM has high measurement accuracy.Cross-experiments are also conducted for x86 and ARM architecture servers.,and experimental results show that the CPU power beha-viors of servers with different architectures are not the same,thus different CPU power modeling methods should be used.

Key words: Energy consumption optimization, Power model, CPU heterogeneity, ARM architecture

中图分类号: 

  • TP393
[1]SCHAGAEV I,KAEGI-TRACHSEL T.Architecture Comparison and Evaluation[M].Springer International Publishing,2016.
[2]AKRAM A.A Study on the Impact of Instruction Set Architectures on Processor's Performance[D].Davis:University of California,Davis,2017.
[3]SHEN J P,LIPASTI M H.Modern Processor Design:Fundamentals of Superscalar Processors [M].Waveland Press,2013.
[4]WANG W.An improved instruction-level power and energymodel for RISC microprocessors[D].Southampton:University of Southampton,2017.
[5]GREENHALGH P.ARM big.LITTLE Processing with ARMCortex-A15 & Cortex-A7[J/OL].https://www.eetimes.com/big-little-processing-with-arm-cortex-a15-cortex-a7.
[6]DENG L Y P,CHEN C S,CAI H L.Research on Timing Analysis Based on Cache Characteristics of ARM Processor[J].Sichuan Ordnance Journal,2015(11):118-121,124.
[7]BLEM E,MENON J,SANKARALINGAM K.Power Strug-gles:Revisiting the RISC vs.CISC Debate on Contemporary ARM and x86 Architectures [C]//2013 IEEE 19thInterna-tional Symposium on High Performance Computer Architecture(HPCA).IEEE,2013:1-12.
[8]AROCA R V,GONALVES L M G.Towards Green Data-Centers:A Comparison of x86 and ARM Architectures Power Efficiency[J].Journal of Parallel and Distributed Computing,2012,72(12):1770-1780.
[9]LI J W,LUO P.Opportunities and Challenges of Server Based on ARM Architecture[J].China New Telecommunications,2020,22(18):47-48.
[10]VASILAKIS E.An instruction level energy characterization of arm processors[R].FORTH-ICS/TR-450,2015.
[11]OBUKHOVA K,ZHURAVSKA I,BURENKO V.Diagnostics of Power Consumption ofa Mobile Device Multi-core Processor with Detail of Each Core Utilization[C]//2020 IEEE 15th International Conference on Advanced Trends in Radioelectro-nics,Telecommunications and Computer Engineering(TCSET).IEEE,2020:368-372.
[12]WALKER M J,DAS A K,MERRETT G V,et al.Run-timePower Estimation for Mobile and Embedded Asymmetric Multi-core CPUs[C]//Hipeac Workshop on Energy Efficiency with Heterogenous Computing.2015.
[13]CHEN K,KILPATRICK P,NIKOLOPOULOS D S,et al.Cross Architectural Power Modelling[C]//2020 20th IEEE/ACM International Symposium on Cluster,Cloud and Internet Computing(CCGRID).IEEE,2020:390-399.
[14]SANKARAN S,SRIDHAR R.Energy Modeling for Mobile Devices Using Performance Counters [C]//2013 IEEE 56th International Midwest Symposium on Circuits and Systems(MWSCAS).IEEE,2013:441-444.
[15]SAGI M,DOAN N A V,RAPP M,et al.A Lightweight Nonli-near Methodology to Accurately Model Multicore Processor Po-wer[J].IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems,2020,39(11):3152-3164.
[16]ZHANG Y,LIU Y,LI Z,et al.Accurate CPU Power Modeling for Multicore Smartphones[J/OL].Microsoft Research,2015.https://www.microsoft.com/en-us/research/publication/accurate-cpu-power-modeling-for-multicore-smartphones.
[17]KATAOKA H,DUOLIKUN D,ENOKIDO T,et al.PowerConsumption and Computation Models of a Server with a Multi-core CPU and Experiments[C]//2015 IEEE 29th International Conference on Advanced Information Networking and Applications Workshops.IEEE,2015:217-222.
[18]HUANG W,LEFURGY C,KUK W,et al.Accurate Fine-grained Processor Power Proxies[C]//2012 45th Annual IEEE/ACM International Symposium on Microarchitecture.IEEE,2012:224-234.
[19]ZHAI Y,ZHANG X,ERANIAN S,et al.Happy:Hyperthread-aware Power Profiling Dynamically[C]//2014 {USENIX} Annual Technical Conference.2014:211-217.
[20]SHEN K,SHRIRAMAN A,DWARKADAS S,et al.PowerContainers:An OS Facility for Fine-Grained Power and Energy Management on Multicore Servers[J].Computer Architecture News,2013,41(1):65-76.
[21]MISHRA S K,PARIDA P P,SAHOO S,et al.Improving Energy Usage in Cloud Computing Using DVFS[M]//Progress in Advanced Computing and Intelligent Engineering.Singapore:Springer,2018:623-632.
[22]LIN W,YU T,GAO C,et al.A Hardware-aware CPU Power Measurement Based on the Power-exponent Function Model for Cloud Servers[J].Information Sciences,2021,547:1045-1065.
[1] 金小敏, 滑文强.
移动云计算中面向能耗优化的资源管理
Energy Optimization Oriented Resource Management in Mobile Cloud Computing
计算机科学, 2020, 47(6): 247-251. https://doi.org/10.11896/jsjkx.190400020
[2] 胡锦天, 王高才, 徐晓桐.
移动边缘计算中具有能耗优化的任务迁移策略
Task Migration Strategy with Energy Optimization in Mobile Edge Computing
计算机科学, 2020, 47(6): 260-265. https://doi.org/10.11896/jsjkx.190400074
[3] 张彭奕, 宋杰.
区块链共识算法效能优化研究进展
Research Advance on Efficiency Optimization of Blockchain Consensus Algorithms
计算机科学, 2020, 47(12): 296-303. https://doi.org/10.11896/jsjkx.200700020
[4] 卢海峰, 顾春华, 罗飞, 丁炜超, 袁野, 任强.
强化学习下能耗优化的虚拟机放置策略
Virtual Machine Placement Strategy with Energy Consumption Optimization under Reinforcement Learning
计算机科学, 2019, 46(9): 291-297. https://doi.org/10.11896/j.issn.1002-137X.2019.09.044
[5] 庄晓照, 万继光, 张艺文, 瞿晓阳.
一种基于新能源驱动的存储系统的能耗优化方案
Energy Consumption Optimization Scheme for New Energy-driven Storage System
计算机科学, 2018, 45(7): 66-72. https://doi.org/10.11896/j.issn.1002-137X.2018.07.010
[6] 黄荣喜, 王淖, 谢天骁, 王高才.
无线网络中具有信道感知的期望能耗最小化策略研究
Study on Channel-aware Expected Energy Consumption Minimization Strategy in Wireless Networks
计算机科学, 2018, 45(10): 130-137. https://doi.org/10.11896/j.issn.1002-137X.2018.10.025
[7] 许慧青,王高才,闵仁江.
一种基于协同缓存的内容中心网络能耗优化策略
Energy-consumption Optimization Strategy Based on Cooperative Caching for Content-centric Network
计算机科学, 2017, 44(8): 76-81. https://doi.org/10.11896/j.issn.1002-137X.2017.08.014
[8] 廖彬,张陶,于炯,国冰磊,刘炎.
基于Spark的MapReduce相似度计算效率优化
Efficiency Optimization Method for MapReduce Similarity Computing Based on Spark
计算机科学, 2017, 44(8): 46-53. https://doi.org/10.11896/j.issn.1002-137X.2017.08.009
[9] 彭颖,王高才,王淖.
移动网络中基于数据到达速率的数据传输能耗优化策略
Energy Consumption Optimization Strategy for Data Transmission Based on Data Arrival Rate in Mobile Networks
计算机科学, 2017, 44(1): 117-122. https://doi.org/10.11896/j.issn.1002-137X.2017.01.023
[10] 王卓薇,程良伦,肖红.
一种基于GPU的高精度体系结构级功耗模型
High-precision Architecture-level Power Model Based on GPU
计算机科学, 2016, 43(11): 30-35. https://doi.org/10.11896/j.issn.1002-137X.2016.11.006
[11] 王科特,王力生,廖新考.
基于多核处理器的K线程低能耗的任务调度优化算法
K-threaded Low Energy-consuming Task Scheduling Optimization Algorithm Based on Multi-core Processors
计算机科学, 2015, 42(2): 18-23. https://doi.org/10.11896/j.issn.1002-137X.2015.02.004
[12] 国冰磊,于 炯,廖 彬,杨德先.
SQL能耗建模及优化研究
Research on SQL Energy Consumption Modeling and Optimization
计算机科学, 2015, 42(10): 202-207.
[13] 肖志娇,明仲,蔡树彬.
基于状态管理的服务器节能策略研究
Study on Energy Optimization of Servers Based on States Management
计算机科学, 2013, 40(4): 22-25.
[14] 刘会英,王韬,赵新杰,周林.
PRESENT相关功耗分析攻击研究
Research on Correlation Power Analysis Attack against PRESENT
计算机科学, 2011, 38(11): 40-42.
[15] 俞莉花,曾国荪.
异构计算中的时间和能耗优化执行方法
Executing Method of Time and Energy Optimization in Heterogeneous Computing
计算机科学, 2011, 38(10): 285-290.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!