Newmark local time stepping on high-performance computing architectures

Handle URI:
http://hdl.handle.net/10754/621888
Title:
Newmark local time stepping on high-performance computing architectures
Authors:
Rietmann, Max; Grote, Marcus; Peter, Daniel ( 0000-0002-3397-5379 ) ; Schenk, Olaf
Abstract:
In multi-scale complex media, finite element meshes often require areas of local refinement, creating small elements that can dramatically reduce the global time-step for wave-propagation problems due to the CFL condition. Local time stepping (LTS) algorithms allow an explicit time-stepping scheme to adapt the time-step to the element size, allowing near-optimal time-steps everywhere in the mesh. We develop an efficient multilevel LTS-Newmark scheme and implement it in a widely used continuous finite element seismic wave-propagation package. In particular, we extend the standard LTS formulation with adaptations to continuous finite element methods that can be implemented very efficiently with very strong element-size contrasts (more than 100×). Capable of running on large CPU and GPU clusters, we present both synthetic validation examples and large scale, realistic application examples to demonstrate the performance and applicability of the method and implementation on thousands of CPU cores and hundreds of GPUs.
KAUST Department:
Extreme Computing Research Center; Physical Sciences and Engineering (PSE) Division
Citation:
Rietmann M, Grote M, Peter D, Schenk O (2016) Newmark local time stepping on high-performance computing architectures. Journal of Computational Physics. Available: http://dx.doi.org/10.1016/j.jcp.2016.11.012.
Publisher:
Elsevier BV
Journal:
Journal of Computational Physics
Issue Date:
25-Nov-2016
DOI:
10.1016/j.jcp.2016.11.012
Type:
Article
ISSN:
0021-9991
Sponsors:
The computational resources and services used in this work were provided by the Swiss National Supercomputing Centre (CSCS). D. Peter and M. Rietmann were supported by the Swiss PASC project “A framework for multi-scale seismic modelling and inversion.” SPECFEM3D_Cartesian is hosted by the Computational Infrastructure for Geodynamics (CIG) which is supported by the National Science Foundation award NSF-0949446.
Additional Links:
http://www.sciencedirect.com/science/article/pii/S0021999116305988
Appears in Collections:
Articles; Physical Sciences and Engineering (PSE) Division; Extreme Computing Research Center

Full metadata record

DC FieldValue Language
dc.contributor.authorRietmann, Maxen
dc.contributor.authorGrote, Marcusen
dc.contributor.authorPeter, Danielen
dc.contributor.authorSchenk, Olafen
dc.date.accessioned2016-11-27T08:37:28Z-
dc.date.available2016-11-27T08:37:28Z-
dc.date.issued2016-11-25en
dc.identifier.citationRietmann M, Grote M, Peter D, Schenk O (2016) Newmark local time stepping on high-performance computing architectures. Journal of Computational Physics. Available: http://dx.doi.org/10.1016/j.jcp.2016.11.012.en
dc.identifier.issn0021-9991en
dc.identifier.doi10.1016/j.jcp.2016.11.012en
dc.identifier.urihttp://hdl.handle.net/10754/621888-
dc.description.abstractIn multi-scale complex media, finite element meshes often require areas of local refinement, creating small elements that can dramatically reduce the global time-step for wave-propagation problems due to the CFL condition. Local time stepping (LTS) algorithms allow an explicit time-stepping scheme to adapt the time-step to the element size, allowing near-optimal time-steps everywhere in the mesh. We develop an efficient multilevel LTS-Newmark scheme and implement it in a widely used continuous finite element seismic wave-propagation package. In particular, we extend the standard LTS formulation with adaptations to continuous finite element methods that can be implemented very efficiently with very strong element-size contrasts (more than 100×). Capable of running on large CPU and GPU clusters, we present both synthetic validation examples and large scale, realistic application examples to demonstrate the performance and applicability of the method and implementation on thousands of CPU cores and hundreds of GPUs.en
dc.description.sponsorshipThe computational resources and services used in this work were provided by the Swiss National Supercomputing Centre (CSCS). D. Peter and M. Rietmann were supported by the Swiss PASC project “A framework for multi-scale seismic modelling and inversion.” SPECFEM3D_Cartesian is hosted by the Computational Infrastructure for Geodynamics (CIG) which is supported by the National Science Foundation award NSF-0949446.en
dc.publisherElsevier BVen
dc.relation.urlhttp://www.sciencedirect.com/science/article/pii/S0021999116305988en
dc.rightsNOTICE: this is the author’s version of a work that was accepted for publication in Journal of Computational Physics. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Journal of Computational Physics, 25 November 2016. DOI: 10.1016/j.jcp.2016.11.012en
dc.subjectSeismic wave propagationen
dc.subjectLocal-time steppingen
dc.subjectMulti-rate time steppingen
dc.titleNewmark local time stepping on high-performance computing architecturesen
dc.typeArticleen
dc.contributor.departmentExtreme Computing Research Centeren
dc.contributor.departmentPhysical Sciences and Engineering (PSE) Divisionen
dc.identifier.journalJournal of Computational Physicsen
dc.eprint.versionPost-printen
dc.contributor.institutionInstitute for Computational Science, Università della Svizzera italiana, Lugano, Switzerlanden
dc.contributor.institutionInstitute of Geophysics, ETH Zurich, Switzerlanden
dc.contributor.institutionDepartment of Mathematics and Computer Science, University of Basel, Switzerlanden
kaust.authorPeter, Danielen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.