Accelerating the conjugate gradient algorithm with gpus in cfd simulations

Hartwig Anzt; Marc Baboulin; Jack Dongarra; Yvan Fournier; Frank Hulsemann; Amal Khabou; Yushan Wang

doi:10.1007/978-3-319-61982-8_5

Accelerating the conjugate gradient algorithm with gpus in cfd simulations

Hartwig Anzt, Marc Baboulin, Jack Dongarra, Yvan Fournier, Frank Hulsemann, Amal Khabou, Yushan Wang

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

This paper illustrates how GPU computing can be used to accelerate computational fluid dynamics (CFD) simulations. For sparse linear systems arising from finite volume discretization, we evaluate and optimize the performance of Conjugate Gradient (CG) routines designed for manycore accelerators and compare against an industrial CPU-based implementation. We also investigate how the recent advances in preconditioning, such as iterative Incomplete Cholesky (IC, as symmetric case of ILU) preconditioning, match the requirements for solving real world problems.

Original language	English
Title of host publication	High Performance Computing for Computational Science
Subtitle of host publication	VECPAR 2016 - 12th International Conference, Revised Selected Papers
Editors	Ines Dutra, Rui Camacho, Jorge Barbosa, Osni Marques
Publisher	Springer Verlag
Pages	35-43
Number of pages	9
ISBN (Print)	9783319619811
DOIs	https://doi.org/10.1007/978-3-319-61982-8_5
State	Published - 2017
Externally published	Yes
Event	12th International Conference on High Performance Computing for Computational Science, VECPAR 2016 - Porto, Portugal Duration: Jun 28 2016 → Jun 30 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10150 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	12th International Conference on High Performance Computing for Computational Science, VECPAR 2016
Country/Territory	Portugal
City	Porto
Period	06/28/16 → 06/30/16

Access to Document

10.1007/978-3-319-61982-8_5

Cite this

Anzt, H., Baboulin, M., Dongarra, J., Fournier, Y., Hulsemann, F., Khabou, A., & Wang, Y. (2017). Accelerating the conjugate gradient algorithm with gpus in cfd simulations. In I. Dutra, R. Camacho, J. Barbosa, & O. Marques (Eds.), High Performance Computing for Computational Science: VECPAR 2016 - 12th International Conference, Revised Selected Papers (pp. 35-43). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10150 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-61982-8_5

Anzt, Hartwig ; Baboulin, Marc ; Dongarra, Jack et al. / Accelerating the conjugate gradient algorithm with gpus in cfd simulations. High Performance Computing for Computational Science: VECPAR 2016 - 12th International Conference, Revised Selected Papers. editor / Ines Dutra ; Rui Camacho ; Jorge Barbosa ; Osni Marques. Springer Verlag, 2017. pp. 35-43 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{599a8de601004d54a1a1c7b083bc9ec4,

title = "Accelerating the conjugate gradient algorithm with gpus in cfd simulations",

abstract = "This paper illustrates how GPU computing can be used to accelerate computational fluid dynamics (CFD) simulations. For sparse linear systems arising from finite volume discretization, we evaluate and optimize the performance of Conjugate Gradient (CG) routines designed for manycore accelerators and compare against an industrial CPU-based implementation. We also investigate how the recent advances in preconditioning, such as iterative Incomplete Cholesky (IC, as symmetric case of ILU) preconditioning, match the requirements for solving real world problems.",

author = "Hartwig Anzt and Marc Baboulin and Jack Dongarra and Yvan Fournier and Frank Hulsemann and Amal Khabou and Yushan Wang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; 12th International Conference on High Performance Computing for Computational Science, VECPAR 2016 ; Conference date: 28-06-2016 Through 30-06-2016",

year = "2017",

doi = "10.1007/978-3-319-61982-8_5",

language = "English",

isbn = "9783319619811",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "35--43",

editor = "Ines Dutra and Rui Camacho and Jorge Barbosa and Osni Marques",

booktitle = "High Performance Computing for Computational Science",

}

Anzt, H, Baboulin, M, Dongarra, J, Fournier, Y, Hulsemann, F, Khabou, A & Wang, Y 2017, Accelerating the conjugate gradient algorithm with gpus in cfd simulations. in I Dutra, R Camacho, J Barbosa & O Marques (eds), High Performance Computing for Computational Science: VECPAR 2016 - 12th International Conference, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10150 LNCS, Springer Verlag, pp. 35-43, 12th International Conference on High Performance Computing for Computational Science, VECPAR 2016, Porto, Portugal, 06/28/16. https://doi.org/10.1007/978-3-319-61982-8_5

Accelerating the conjugate gradient algorithm with gpus in cfd simulations. / Anzt, Hartwig; Baboulin, Marc; Dongarra, Jack et al.
High Performance Computing for Computational Science: VECPAR 2016 - 12th International Conference, Revised Selected Papers. ed. / Ines Dutra; Rui Camacho; Jorge Barbosa; Osni Marques. Springer Verlag, 2017. p. 35-43 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10150 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Accelerating the conjugate gradient algorithm with gpus in cfd simulations

AU - Anzt, Hartwig

AU - Baboulin, Marc

AU - Dongarra, Jack

AU - Fournier, Yvan

AU - Hulsemann, Frank

AU - Khabou, Amal

AU - Wang, Yushan

N1 - Publisher Copyright: © Springer International Publishing AG 2017.

PY - 2017

Y1 - 2017

N2 - This paper illustrates how GPU computing can be used to accelerate computational fluid dynamics (CFD) simulations. For sparse linear systems arising from finite volume discretization, we evaluate and optimize the performance of Conjugate Gradient (CG) routines designed for manycore accelerators and compare against an industrial CPU-based implementation. We also investigate how the recent advances in preconditioning, such as iterative Incomplete Cholesky (IC, as symmetric case of ILU) preconditioning, match the requirements for solving real world problems.

AB - This paper illustrates how GPU computing can be used to accelerate computational fluid dynamics (CFD) simulations. For sparse linear systems arising from finite volume discretization, we evaluate and optimize the performance of Conjugate Gradient (CG) routines designed for manycore accelerators and compare against an industrial CPU-based implementation. We also investigate how the recent advances in preconditioning, such as iterative Incomplete Cholesky (IC, as symmetric case of ILU) preconditioning, match the requirements for solving real world problems.

UR - http://www.scopus.com/inward/record.url?scp=85026788315&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-61982-8_5

DO - 10.1007/978-3-319-61982-8_5

M3 - Conference contribution

AN - SCOPUS:85026788315

SN - 9783319619811

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 35

EP - 43

BT - High Performance Computing for Computational Science

A2 - Dutra, Ines

A2 - Camacho, Rui

A2 - Barbosa, Jorge

A2 - Marques, Osni

PB - Springer Verlag

T2 - 12th International Conference on High Performance Computing for Computational Science, VECPAR 2016

Y2 - 28 June 2016 through 30 June 2016

ER -

Anzt H, Baboulin M, Dongarra J, Fournier Y, Hulsemann F, Khabou A et al. Accelerating the conjugate gradient algorithm with gpus in cfd simulations. In Dutra I, Camacho R, Barbosa J, Marques O, editors, High Performance Computing for Computational Science: VECPAR 2016 - 12th International Conference, Revised Selected Papers. Springer Verlag. 2017. p. 35-43. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-61982-8_5

Accelerating the conjugate gradient algorithm with gpus in cfd simulations

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this