{"title":"DVFS Considering Spatial Correlation Timing and Process-Voltage-Temperature Variations","authors":"Tung-Liang Lin, Sao-Jie Chen","doi":"10.1109/socc49529.2020.9524768","DOIUrl":null,"url":null,"abstract":"A novel scheme, spatially-correlated Design Dependent Critical-Path Monitor (DDCPM), is proposed, which can provide valuable references in deriving application-specific, process- and temperature-aware DVFS for aggressive power saving during runtime. Such DDCPM utilizes its unique spatial correlation feature and real-time sampling techniques to precisely sense the unexpected behavior introduced by over-scaled voltage under the operating conditions with random and mutually dependent Process-Voltage-Temperature (PVT) variations in each individual chip. Our experimental results obtained in two IPs implemented in TSMC 28 nm process node respectively show average step-wise 7.80% and 8.19% power could be reduced at a smaller granular level of voltage scaling, which corresponding maximum power reductions, 55.6% and 57.5% in Typical Corner could be finally achieved.","PeriodicalId":114740,"journal":{"name":"2020 IEEE 33rd International System-on-Chip Conference (SOCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2020-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 IEEE 33rd International System-on-Chip Conference (SOCC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/socc49529.2020.9524768","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
A novel scheme, spatially-correlated Design Dependent Critical-Path Monitor (DDCPM), is proposed, which can provide valuable references in deriving application-specific, process- and temperature-aware DVFS for aggressive power saving during runtime. Such DDCPM utilizes its unique spatial correlation feature and real-time sampling techniques to precisely sense the unexpected behavior introduced by over-scaled voltage under the operating conditions with random and mutually dependent Process-Voltage-Temperature (PVT) variations in each individual chip. Our experimental results obtained in two IPs implemented in TSMC 28 nm process node respectively show average step-wise 7.80% and 8.19% power could be reduced at a smaller granular level of voltage scaling, which corresponding maximum power reductions, 55.6% and 57.5% in Typical Corner could be finally achieved.