Combine dynamic time-slice scaling with DVFS for coordinating thermal and fairness on CPU

Gangyong Jia, Guangjie Han
{"title":"Combine dynamic time-slice scaling with DVFS for coordinating thermal and fairness on CPU","authors":"Gangyong Jia, Guangjie Han","doi":"10.1109/ComComAp.2014.7017169","DOIUrl":null,"url":null,"abstract":"As power density increases with high technology, the high temperature has threatened the system performance, reliability and even system safety. Develop a thermal management for reducing temperature, but without disturbing fairness, is becoming more and more important. Therefore, in this paper, we propose a DTS-DVFS management which combines Dynamic Time-Slice Scaling (DTS) with Dynamic Voltage and Frequency Scaling (DVFS). Through fine-grained thermal characterization based on task behavior, dynamically determine both Time-slice Scaling factor (TSF) and Voltage and Frequency Scaling factor (VSF) for each task on real-time which not only reduces temperature but also retains fairness. Besides scaling both time-slice and voltage and frequency for each task according to TSF and VSF respectively, combining alternative scheduling scheme based on boosting thermal model which predicts the temperature of the chip. Through our experiments in real machine, results demonstrate our proposed policy can reduce the chip average and peak temperature maximums by 4.2°C and 2.9°C with negligible fairness loss.","PeriodicalId":422906,"journal":{"name":"2014 IEEE Computers, Communications and IT Applications Conference","volume":"32 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2014-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2014 IEEE Computers, Communications and IT Applications Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ComComAp.2014.7017169","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1

Abstract

As power density increases with high technology, the high temperature has threatened the system performance, reliability and even system safety. Develop a thermal management for reducing temperature, but without disturbing fairness, is becoming more and more important. Therefore, in this paper, we propose a DTS-DVFS management which combines Dynamic Time-Slice Scaling (DTS) with Dynamic Voltage and Frequency Scaling (DVFS). Through fine-grained thermal characterization based on task behavior, dynamically determine both Time-slice Scaling factor (TSF) and Voltage and Frequency Scaling factor (VSF) for each task on real-time which not only reduces temperature but also retains fairness. Besides scaling both time-slice and voltage and frequency for each task according to TSF and VSF respectively, combining alternative scheduling scheme based on boosting thermal model which predicts the temperature of the chip. Through our experiments in real machine, results demonstrate our proposed policy can reduce the chip average and peak temperature maximums by 4.2°C and 2.9°C with negligible fairness loss.
将动态时间片缩放与DVFS相结合,以协调CPU的散热和公平性
随着技术含量的提高,功率密度不断增大,高温对系统的性能、可靠性甚至系统安全都构成了威胁。开发一种既能降低温度又不影响公平性的热管理方法,显得越来越重要。因此,本文提出了一种将动态时间片缩放(DTS)和动态电压频率缩放(DVFS)相结合的DTS-DVFS管理方法。通过基于任务行为的细粒度热表征,实时动态确定每个任务的时间片比例因子(TSF)和电压频率比例因子(VSF),既降低了温度,又保持了公平性。除了分别根据TSF和VSF对每个任务进行时间片缩放和电压频率缩放外,还结合了基于boost热模型预测芯片温度的备选调度方案。通过我们在真实机器上的实验,结果表明我们提出的策略可以将芯片的平均温度和峰值温度分别降低4.2°C和2.9°C,而公平性损失可以忽略不计。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信