###

DOI:

计算机系统应用英文版:2013,22(10):1-7,38

View/Add Comment 过刊浏览高级检索 HTML

←前一篇 | 后一篇→

码上扫一扫！

下载全文

基于任务行为分析的DVFS机制

陈云^1,2, 贾刚勇^1,2, 李曦^1,2, 张海鹏^1,2

(1.中国科学技术大学计算机科学与技术学院, 合肥 230026;2.中国科学技术大学苏州研究院, 苏州 215028)

Task Behavior Based on DVFS Mechanism

CHEN Yun^1,2, JIA Gang-Yong^1,2, LI Xi^1,2, ZHANG Hai-Peng^1,2

(1.Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China;2.Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China)

摘要

图/表

参考文献

相似文献

本文已被：浏览 1857次下载 4707次
Received:March 25, 2013 Revised:May 06, 2013

中文摘要: 动态电压频率缩放(DVFS)技术是当前最有效的功耗调节手段之一. 本文首先分析现有DVFS技术存在的不足, 指出限制DVFS技术高效运用的核心因素;基于现有低效的方式我们提出一种基于任务行为分析的DVFS机制(TC-DVFS).其具有三个层次: 一、采集任务的系统调用信息;二、识别任务的关键系统调用, 并以关键系统调用刻画任务行为;三、根据任务行为构建特征库, 并以任务的特征库来指导DVFS. 我们将TC-DVFS添加到linux内核中, 并在intel-core2处理器平台上对不同类型的应用任务进行性能与功耗测试. 结果显示TC-DVFS总体获得10%的性能提升, 并降低5%调频失效率和5%的系统能耗.

中文关键词: DVFS 任务行为分析功耗优化系统调用

Abstract:The dynamic voltage and frequency scaling (DVFS) is one of the most effective mechanism for power management. Firstly, we point out the most shortcomings of the current DVFS mechanisms after detailly analyzing;so we propose a task behavior based DVFS mechanism (TC-DVFS). TC-DVFS has three levels: First, gather the system call information of the task. Second, distinguish the critical system calls and characterize task behavior with them;Third, construct a feature database based on task behavior and guide DVFS according the database. Adding TC-DVFS into linux kernel, experimental results show TC-DVFS reduce 5% miss rate of the correct frequency scaling and 5% power of the system, and also improve 10% performance of the system.

keywords: DVFS task behavior analysis power system call

文章编号： 中图分类号： 文献标志码：

基金项目:国家自然基金(61272131,61202053);江苏省自然科学基金(SBK201240198);江苏江苏省产学研前瞻性联合研究项目(BY2009128)

Author Name	Affiliation
CHEN Yun	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China
JIA Gang-Yong	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China
LI Xi	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China
ZHANG Hai-Peng	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China

Author Name	Affiliation
CHEN Yun	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China
JIA Gang-Yong	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China
LI Xi	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China
ZHANG Hai-Peng	Institute of Computer Science and Technology, The University of Science and Technology of China, Hefei 230026, China Suzhou Institute for Advanced Study, The University of Science and Technology of China, Suzhou 215028, China

引用文本：
陈云,贾刚勇,李曦,张海鹏.基于任务行为分析的DVFS机制.计算机系统应用,2013,22(10):1-7,38
CHEN Yun,JIA Gang-Yong,LI Xi,ZHANG Hai-Peng.Task Behavior Based on DVFS Mechanism.COMPUTER SYSTEMS APPLICATIONS,2013,22(10):1-7,38