Methods of parallel computation applied on granular simulations

Methods of parallel computation applied on granular simulations

Every year, parallel computing has becoming cheaper and more accessible. As consequence, applications were spreading over all research areas. Granular materials is a promising area for parallel computing. To prove this statement we study the impact of parallel computing in simulations of the BNE (Br...

Full description

Bibliographic Details
Main Authors: Martins Gustavo H. B., Atman Allbens P. F.
Format: Article
Language:English
Published: EDP Sciences 2017-01-01
Series:EPJ Web of Conferences
Online Access:https://doi.org/10.1051/epjconf/201714015024
id doaj-45b8185d17074e92bce5d641e29eb69e
record_format Article
spelling doaj-45b8185d17074e92bce5d641e29eb69e2021-08-02T06:04:23ZengEDP SciencesEPJ Web of Conferences2100-014X2017-01-011401502410.1051/epjconf/201714015024epjconf162450Methods of parallel computation applied on granular simulationsMartins Gustavo H. B.0Atman Allbens P. F.Programa de Pós-Graduação em Modelagem Matemática e Computacional - PPG-MMC, Centro Federal de Educação Tecnológica de Minas Gerais, CEFET–MGEvery year, parallel computing has becoming cheaper and more accessible. As consequence, applications were spreading over all research areas. Granular materials is a promising area for parallel computing. To prove this statement we study the impact of parallel computing in simulations of the BNE (Brazil Nut Effect). This property is due the remarkable arising of an intruder confined to a granular media when vertically shaken against gravity. By means of DEM (Discrete Element Methods) simulations, we study the code performance testing different methods to improve clock time. A comparison between serial and parallel algorithms, using OpenMP® is also shown. The best improvement was obtained by optimizing the function that find contacts using Verlet’s cells.https://doi.org/10.1051/epjconf/201714015024
collection DOAJ
language English
format Article
sources DOAJ
author Martins Gustavo H. B.
Atman Allbens P. F.
spellingShingle Martins Gustavo H. B.
Atman Allbens P. F.
Methods of parallel computation applied on granular simulations
EPJ Web of Conferences
author_facet Martins Gustavo H. B.
Atman Allbens P. F.
author_sort Martins Gustavo H. B.
title Methods of parallel computation applied on granular simulations
title_short Methods of parallel computation applied on granular simulations
title_full Methods of parallel computation applied on granular simulations
title_fullStr Methods of parallel computation applied on granular simulations
title_full_unstemmed Methods of parallel computation applied on granular simulations
title_sort methods of parallel computation applied on granular simulations
publisher EDP Sciences
series EPJ Web of Conferences
issn 2100-014X
publishDate 2017-01-01
description Every year, parallel computing has becoming cheaper and more accessible. As consequence, applications were spreading over all research areas. Granular materials is a promising area for parallel computing. To prove this statement we study the impact of parallel computing in simulations of the BNE (Brazil Nut Effect). This property is due the remarkable arising of an intruder confined to a granular media when vertically shaken against gravity. By means of DEM (Discrete Element Methods) simulations, we study the code performance testing different methods to improve clock time. A comparison between serial and parallel algorithms, using OpenMP® is also shown. The best improvement was obtained by optimizing the function that find contacts using Verlet’s cells.
url https://doi.org/10.1051/epjconf/201714015024
work_keys_str_mv AT martinsgustavohb methodsofparallelcomputationappliedongranularsimulations
AT atmanallbenspf methodsofparallelcomputationappliedongranularsimulations
_version_ 1721240610747187200

Similar Items