{"title":"Parallel implementation of 3D-EFIT simulation using MPI library for elastic waves","authors":"P. Ampha, A. Ruengwaree","doi":"10.1109/ECTICON.2016.7561483","DOIUrl":null,"url":null,"abstract":"In present paper demonstrated the implementation of the three dimensional time-domain wavefield computation technique for a large and realistic modeling which called Elastodynamic Finite Integration Technique (EFIT). The disadvantage of the time-domain computation are time consumptions and memory requirements. In this work, we apply the portable message passing interface standard (MPI) for the domain decomposition and the communication between processes. For the outer boundary conditions, we have used convolutional perfectly matched layer (C-PML) as open boundaries to terminate the EFIT lattices. For the numerical simulation example, we performed 2 numerical models with size of 1000×1000×600 cells with 16 processors on cluster computer nodes. The MPI based 3D-EFIT code shows good performance on speed up, efficiency, and elapsed time.","PeriodicalId":200661,"journal":{"name":"2016 13th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 13th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ECTICON.2016.7561483","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
In present paper demonstrated the implementation of the three dimensional time-domain wavefield computation technique for a large and realistic modeling which called Elastodynamic Finite Integration Technique (EFIT). The disadvantage of the time-domain computation are time consumptions and memory requirements. In this work, we apply the portable message passing interface standard (MPI) for the domain decomposition and the communication between processes. For the outer boundary conditions, we have used convolutional perfectly matched layer (C-PML) as open boundaries to terminate the EFIT lattices. For the numerical simulation example, we performed 2 numerical models with size of 1000×1000×600 cells with 16 processors on cluster computer nodes. The MPI based 3D-EFIT code shows good performance on speed up, efficiency, and elapsed time.