Ramesha Rehman, Mashood Ul Haq Chishti, Hamza Yamin
{"title":"Efficient GPU Power Management through Advanced Framework Utilizing Optimization Algorithms","authors":"Ramesha Rehman, Mashood Ul Haq Chishti, Hamza Yamin","doi":"10.56415/csjm.v32.08","DOIUrl":null,"url":null,"abstract":"The rapid rise in power usage by GPUs due to advances in machine and deep learning has led to an increase in power consumption of GPUs in Deep Learning workloads. To address this issue, a novel research project focuses on integrating Particle Swarm Optimization into a model training optimization framework to effectively reduce GPU power consumption during machine learning and deep learning training workloads. By utilizing the Particle Swarm Optimization (PSO)\\protect\\hyperlink{b1}{{[}1{]}} algorithm within the proposed framework, we show the effectiveness of PSO in creating a more efficient power management strategy while also maintaining the performance. Upon evaluation of the proposed framework, it shows a reduction of 15.8\\% to 75.8\\% in power consumption across multiple workloads, with little to no performance loss.","PeriodicalId":42293,"journal":{"name":"Computer Science Journal of Moldova","volume":null,"pages":null},"PeriodicalIF":0.2000,"publicationDate":"2024-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computer Science Journal of Moldova","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.56415/csjm.v32.08","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"COMPUTER SCIENCE, THEORY & METHODS","Score":null,"Total":0}
引用次数: 0
Abstract
The rapid rise in power usage by GPUs due to advances in machine and deep learning has led to an increase in power consumption of GPUs in Deep Learning workloads. To address this issue, a novel research project focuses on integrating Particle Swarm Optimization into a model training optimization framework to effectively reduce GPU power consumption during machine learning and deep learning training workloads. By utilizing the Particle Swarm Optimization (PSO)\protect\hyperlink{b1}{{[}1{]}} algorithm within the proposed framework, we show the effectiveness of PSO in creating a more efficient power management strategy while also maintaining the performance. Upon evaluation of the proposed framework, it shows a reduction of 15.8\% to 75.8\% in power consumption across multiple workloads, with little to no performance loss.