PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS

Bibliographic Details
Main Author: Martins de Paula, Lauro Cássio; Instituto de Informática - Universidade Federal de Goiás
Publication Date: 2013
Language: por
Source: Revista de Sistemas e Computação
Download full: https://revistas.unifacs.br/index.php/rsc/article/view/2832
Summary: This paper presents a parallel implementation of the Hybrid Bi-Conjugate Gradient Stabilized (BiCGStab(2)) iterative method in Graphics Processing Unit (GPU) for solution of large and sparse linear systems. This implementation uses the CUDA-Matlab integration, in which the method operations are performed in a GPU cores using Matlab built-in functions. The goal is to show that the exploitation of parallelism by using this new technology can provide a significant computational performance. For the validation of the work we compared the proposed implementation with a BiCGStab(2) sequential and parallelized implementation in the C and CUDA-C languages, respectively. The results showed that the proposed implementation is more efficient and can be indispensable for simulations being carried out with quality and in a timely manner. The gains in computational efficiency were, respectively, 76x and 6x compared to the implementation in C and CUDA-C.
id UNIF-1_b0df6d2e98d9aeee90af4e48ef71cf92
oai_identifier_str oai:ojs.200.223.74.126:article/2832
network_acronym_str UNIF-1
network_name_str Revista de Sistemas e Computação
repository_id_str
spelling PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMSIMPLEMENTAÇÃO PARALELA DO MÉTODO BICGSTAB(2) EM GPU USANDO CUDA E MATLAB PARA SOLUÇÃO DE SISTEMAS LINEARESMatlab; GPU; CUDA; BiCGStab(2Matlab; GPU; CUDA; BiCGStab(2)This paper presents a parallel implementation of the Hybrid Bi-Conjugate Gradient Stabilized (BiCGStab(2)) iterative method in Graphics Processing Unit (GPU) for solution of large and sparse linear systems. This implementation uses the CUDA-Matlab integration, in which the method operations are performed in a GPU cores using Matlab built-in functions. The goal is to show that the exploitation of parallelism by using this new technology can provide a significant computational performance. For the validation of the work we compared the proposed implementation with a BiCGStab(2) sequential and parallelized implementation in the C and CUDA-C languages, respectively. The results showed that the proposed implementation is more efficient and can be indispensable for simulations being carried out with quality and in a timely manner. The gains in computational efficiency were, respectively, 76x and 6x compared to the implementation in C and CUDA-C.Este artigo apresenta uma implementação paralela do método iterativo Gradiente Bi-Conjugado Estabilizado Híbrido (BiCGStab(2)) em Graphics Processing Unit (GPU) para solução de sistemas lineares grandes e esparsos. Tal implementação faz uso da integração CUDA-Matlab, em que as operações do método são executadas nos núcleos de uma GPU por meio de funções padrão do Matlab. O objetivo é mostrar que a exploração de paralelismo utilizando essa nova tecnologia pode fornecer um desempenho computacional significante. Para a validação do trabalho, comparou-se a implementação proposta com uma implementação sequencial e outra paralelizada do BiCGStab(2) nas linguagens C e CUDA-C, respectivamente. Os resultados mostraram que a implementação proposta é mais eficiente e pode ser indispensável para que simulações sejam realizadas com qualidade e em um tempo hábil. Os ganhos de eficiência computacional foram de, respectivamente, 76x e 6x em relação à implementação em C e CUDA-C.Revista de Sistemas e Computação - RSCRevistade Sistemas y ComputaciónCAPESMartins de Paula, Lauro Cássio; Instituto de Informática - Universidade Federal de Goiás2013-12-24Artigo Avaliado pelos Paresinfo:eu-repo/semantics/publishedVersionapplication/pdfhttps://revistas.unifacs.br/index.php/rsc/article/view/283210.36558/rsc.v3i2.2832Revista de Sistemas e Computação - RSC; v. 3, n. 2 (2013)Revistade Sistemas y Computación; v. 3, n. 2 (2013)reponame:Revista de Sistemas e Computaçãoinstname:Universidade Salvador (UNIFACS)instacron:UNIFACSporinfo:eu-repo/semantics/openAccess2013-12-24T04:28:33Zoai:ojs.200.223.74.126:article/2832Revistahttps://revistas.unifacs.br/index.php/rscPRIhttps://revistas.unifacs.br/index.php/rsc/oaipaulo.caetano@unifacs.br || unifacs@nexodoc.com.br2237-29032237-2903opendoar:2013-12-24T04:28:33Revista de Sistemas e Computação - Universidade Salvador (UNIFACS)false
dc.title.none.fl_str_mv PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
IMPLEMENTAÇÃO PARALELA DO MÉTODO BICGSTAB(2) EM GPU USANDO CUDA E MATLAB PARA SOLUÇÃO DE SISTEMAS LINEARES
title PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
spellingShingle PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
Martins de Paula, Lauro Cássio; Instituto de Informática - Universidade Federal de Goiás
Matlab; GPU; CUDA; BiCGStab(2
Matlab; GPU; CUDA; BiCGStab(2)
title_short PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
title_full PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
title_fullStr PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
title_full_unstemmed PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
title_sort PARALLEL IMPLEMENTATION OF THE BICGSTAB(2) METHOD IN GPU USING CUDA AND MATLAB FOR SOLUTION OF LINEAR SYSTEMS
author Martins de Paula, Lauro Cássio; Instituto de Informática - Universidade Federal de Goiás
author_facet Martins de Paula, Lauro Cássio; Instituto de Informática - Universidade Federal de Goiás
author_role author
dc.contributor.none.fl_str_mv
CAPES
dc.contributor.author.fl_str_mv Martins de Paula, Lauro Cássio; Instituto de Informática - Universidade Federal de Goiás
dc.subject.por.fl_str_mv Matlab; GPU; CUDA; BiCGStab(2
Matlab; GPU; CUDA; BiCGStab(2)
topic Matlab; GPU; CUDA; BiCGStab(2
Matlab; GPU; CUDA; BiCGStab(2)
description This paper presents a parallel implementation of the Hybrid Bi-Conjugate Gradient Stabilized (BiCGStab(2)) iterative method in Graphics Processing Unit (GPU) for solution of large and sparse linear systems. This implementation uses the CUDA-Matlab integration, in which the method operations are performed in a GPU cores using Matlab built-in functions. The goal is to show that the exploitation of parallelism by using this new technology can provide a significant computational performance. For the validation of the work we compared the proposed implementation with a BiCGStab(2) sequential and parallelized implementation in the C and CUDA-C languages, respectively. The results showed that the proposed implementation is more efficient and can be indispensable for simulations being carried out with quality and in a timely manner. The gains in computational efficiency were, respectively, 76x and 6x compared to the implementation in C and CUDA-C.
publishDate 2013
dc.date.none.fl_str_mv 2013-12-24
dc.type.driver.fl_str_mv Artigo Avaliado pelos Pares
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://revistas.unifacs.br/index.php/rsc/article/view/2832
10.36558/rsc.v3i2.2832
url https://revistas.unifacs.br/index.php/rsc/article/view/2832
identifier_str_mv 10.36558/rsc.v3i2.2832
dc.language.iso.fl_str_mv por
language por
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Revista de Sistemas e Computação - RSC
Revistade Sistemas y Computación
publisher.none.fl_str_mv Revista de Sistemas e Computação - RSC
Revistade Sistemas y Computación
dc.source.none.fl_str_mv Revista de Sistemas e Computação - RSC; v. 3, n. 2 (2013)
Revistade Sistemas y Computación; v. 3, n. 2 (2013)
reponame:Revista de Sistemas e Computação
instname:Universidade Salvador (UNIFACS)
instacron:UNIFACS
instname_str Universidade Salvador (UNIFACS)
instacron_str UNIFACS
institution UNIFACS
reponame_str Revista de Sistemas e Computação
collection Revista de Sistemas e Computação
repository.name.fl_str_mv Revista de Sistemas e Computação - Universidade Salvador (UNIFACS)
repository.mail.fl_str_mv paulo.caetano@unifacs.br || unifacs@nexodoc.com.br
_version_ 1833830804459880448

Similar Items