L. Gasparini, J. Rodrigues, C. Conopoima, D. A. Augusto, Michael Souza, L. M. Carvalho, P. Goldfeld, João Paulo Ramirez, J. Panetta
{"title":"A Linear Solver Framework for Flow and Geomechanics Reservoir Simulation","authors":"L. Gasparini, J. Rodrigues, C. Conopoima, D. A. Augusto, Michael Souza, L. M. Carvalho, P. Goldfeld, João Paulo Ramirez, J. Panetta","doi":"10.1109/IPDPSW.2019.00119","DOIUrl":null,"url":null,"abstract":"This paper describes a parallel solver framework focused on flow and geomechanics reservoir simulation applications. It has been designed to run efficiently on a wide range of target platforms, from desktop workstations to heterogeneous clusters of multicore nodes, with or without GPUs, using a framework for distributed matrices and vectors based on a two-tier hierarchical architecture. Results show good parallel scalability on clusters of multicore nodes. Comparisons with the PETSc library indicate it is competitive with the best available tools. Preliminary tests indicate good speedups and parallel scalability also on multiple GPUs.","PeriodicalId":292054,"journal":{"name":"2019 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IPDPSW.2019.00119","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 3
Abstract
This paper describes a parallel solver framework focused on flow and geomechanics reservoir simulation applications. It has been designed to run efficiently on a wide range of target platforms, from desktop workstations to heterogeneous clusters of multicore nodes, with or without GPUs, using a framework for distributed matrices and vectors based on a two-tier hierarchical architecture. Results show good parallel scalability on clusters of multicore nodes. Comparisons with the PETSc library indicate it is competitive with the best available tools. Preliminary tests indicate good speedups and parallel scalability also on multiple GPUs.