{"title":"A novel scheduling-based CAD methodology for exploring the design space of ASICs for low power","authors":"A. Kumar, M. Bayoumi","doi":"10.1109/ASIC.1998.722814","DOIUrl":null,"url":null,"abstract":"This paper describes a novel approach to scheduling with multiple supply voltages in the high-level synthesis of ASICs. In a significant shift from the existing scheduling algorithms for multiple voltages, the proposed approach considers, identifies, and exploits the maximal parallelism available in an initial schedule, and applies a modified stochastic evolution mechanism to iteratively improve, or re-schedule, the previously obtained best-schedule to reduce the maximal power consumption of function-units. Based on simulation and evaluation of the proposed approach (using standard benchmarks), it is observed that a power savings of up to 80% is possible when three supply voltage levels, 5 V, 3.3 V, and 2.4 V are considered. In addition to scheduling for low power, the proposed methodology can serve as a vital guiding tool to a designer for studying the efficacy of different design choices before a final design option is selected.","PeriodicalId":104431,"journal":{"name":"Proceedings Eleventh Annual IEEE International ASIC Conference (Cat. No.98TH8372)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1998-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"4","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings Eleventh Annual IEEE International ASIC Conference (Cat. No.98TH8372)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ASIC.1998.722814","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 4
Abstract
This paper describes a novel approach to scheduling with multiple supply voltages in the high-level synthesis of ASICs. In a significant shift from the existing scheduling algorithms for multiple voltages, the proposed approach considers, identifies, and exploits the maximal parallelism available in an initial schedule, and applies a modified stochastic evolution mechanism to iteratively improve, or re-schedule, the previously obtained best-schedule to reduce the maximal power consumption of function-units. Based on simulation and evaluation of the proposed approach (using standard benchmarks), it is observed that a power savings of up to 80% is possible when three supply voltage levels, 5 V, 3.3 V, and 2.4 V are considered. In addition to scheduling for low power, the proposed methodology can serve as a vital guiding tool to a designer for studying the efficacy of different design choices before a final design option is selected.