{"title":"复杂物体电磁建模的mpi加速多求解算法","authors":"Jian Guan, Su Yan, Kedi Zhang, Jianming Jin","doi":"10.1109/APUSNCURSINRSM.2017.8072709","DOIUrl":null,"url":null,"abstract":"A multi-solver (MS) algorithm is accelerated on distributed computing systems to simulate the electromagnetic scattering from large and complex objects. In this algorithm, the targeted object with its background is decomposed into multiple subdomains which are modeled by either the finite element method or the method of moments. Different from conventional parallelized domain decomposition methods, where the subdomains are handled in parallel, the MS algorithm is parallelized by applying the parallelization to individual subdomains. The employed multilevel fast multipole algorithm is also parallelized to enable computation on many processors. Numerical examples are given to show the parallel efficiency of the proposed strategy.","PeriodicalId":264754,"journal":{"name":"2017 IEEE International Symposium on Antennas and Propagation & USNC/URSI National Radio Science Meeting","volume":"19 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2017-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An MPI-accelerated multi-solver algorithm for electromagnetic modeling of complex objects\",\"authors\":\"Jian Guan, Su Yan, Kedi Zhang, Jianming Jin\",\"doi\":\"10.1109/APUSNCURSINRSM.2017.8072709\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"A multi-solver (MS) algorithm is accelerated on distributed computing systems to simulate the electromagnetic scattering from large and complex objects. In this algorithm, the targeted object with its background is decomposed into multiple subdomains which are modeled by either the finite element method or the method of moments. Different from conventional parallelized domain decomposition methods, where the subdomains are handled in parallel, the MS algorithm is parallelized by applying the parallelization to individual subdomains. The employed multilevel fast multipole algorithm is also parallelized to enable computation on many processors. Numerical examples are given to show the parallel efficiency of the proposed strategy.\",\"PeriodicalId\":264754,\"journal\":{\"name\":\"2017 IEEE International Symposium on Antennas and Propagation & USNC/URSI National Radio Science Meeting\",\"volume\":\"19 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2017-10-18\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2017 IEEE International Symposium on Antennas and Propagation & USNC/URSI National Radio Science Meeting\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/APUSNCURSINRSM.2017.8072709\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2017 IEEE International Symposium on Antennas and Propagation & USNC/URSI National Radio Science Meeting","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/APUSNCURSINRSM.2017.8072709","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
An MPI-accelerated multi-solver algorithm for electromagnetic modeling of complex objects
A multi-solver (MS) algorithm is accelerated on distributed computing systems to simulate the electromagnetic scattering from large and complex objects. In this algorithm, the targeted object with its background is decomposed into multiple subdomains which are modeled by either the finite element method or the method of moments. Different from conventional parallelized domain decomposition methods, where the subdomains are handled in parallel, the MS algorithm is parallelized by applying the parallelization to individual subdomains. The employed multilevel fast multipole algorithm is also parallelized to enable computation on many processors. Numerical examples are given to show the parallel efficiency of the proposed strategy.