{"title":"Special Aspects of The Use of Parallel Computing in Modeling the Nonlinear Dynamics of Switching Voltage Converters","authors":"A. Andriyanov","doi":"10.1109/RusAutoCon52004.2021.9537454","DOIUrl":null,"url":null,"abstract":"The article deals with parallel computing in the study of the nonlinear dynamics of pulse-width converters. An algorithm for calculating two-parameter diagrams is presented on the example of dynamic mode maps, which implies the use of a computing cluster. A method for calculating the optimal computational load for heterogeneous cluster nodes consisting of computers with different performance has been developed. The routines of the software that implement the specified algorithm have been developed and the efficiency in carrying out scientific calculations is shown. It is shown that increasing the number of nodes in a cluster has limited efficiency and there is an optimal number of them that ensures maximum cluster performance.","PeriodicalId":106150,"journal":{"name":"2021 International Russian Automation Conference (RusAutoCon)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 International Russian Automation Conference (RusAutoCon)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/RusAutoCon52004.2021.9537454","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
The article deals with parallel computing in the study of the nonlinear dynamics of pulse-width converters. An algorithm for calculating two-parameter diagrams is presented on the example of dynamic mode maps, which implies the use of a computing cluster. A method for calculating the optimal computational load for heterogeneous cluster nodes consisting of computers with different performance has been developed. The routines of the software that implement the specified algorithm have been developed and the efficiency in carrying out scientific calculations is shown. It is shown that increasing the number of nodes in a cluster has limited efficiency and there is an optimal number of them that ensures maximum cluster performance.