Self-adaptive multiprecision preconditioners on multicore and Manycore architectures

Hartwig Anzt; Dimitar Lukarski; Stanimire Tomov; Jack Dongarra

doi:10.1007/978-3-319-17353-5_10

Self-adaptive multiprecision preconditioners on multicore and Manycore architectures

Hartwig Anzt, Dimitar Lukarski, Stanimire Tomov, Jack Dongarra

Computer Science and Mathematics Div

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

1 Scopus citations

Abstract

Based on the premise that preconditioners needed for scientific computing are not only required to be robust in the numerical sense, but also scalable for up to thousands of light-weight cores, we argue that this two-fold goal is achieved for the recently developed self-adaptive multi-elimination preconditioner. For this purpose, we revise the underlying idea and analyze the performance of implementations realized in the PARALUTION and MAGMA open-source software libraries on GPU architectures (using either CUDA or OpenCL), Intel’s Many Integrated Core Architecture, and Intel’s Sandy Bridge processor. The comparison with other well-established preconditioners like multi-coloured Gauss- Seidel, ILU(0) and multi-colored ILU(0), shows that the twofold goal of a numerically stable cross-platform performant algorithm is achieved.

Original language	English
Title of host publication	High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers
Editors	Osni Marques, Michel Dayde, Kengo Nakajima
Publisher	Springer Verlag
Pages	115-123
Number of pages	9
ISBN (Print)	9783319173528
DOIs	https://doi.org/10.1007/978-3-319-17353-5_10
State	Published - 2015
Event	11th International Conference on High Performance Computing for Computational Science, VECPAR 2014 - Eugene, United States Duration: Jun 30 2014 → Jul 3 2014

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	8969
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th International Conference on High Performance Computing for Computational Science, VECPAR 2014
Country/Territory	United States
City	Eugene
Period	06/30/14 → 07/3/14

Funding

This work has been supported by the Linnaeus centre of excellence UPMARC, Uppsala Programming for Multicore Architectures Research Center, the Russian Scientific Fund (Agreement N14-11-00190), DOE grant #DE-SC0010042, NVIDIA, and the NSF grant # ACI-1339822.

Access to Document

10.1007/978-3-319-17353-5_10

Cite this

Anzt, H., Lukarski, D., Tomov, S., & Dongarra, J. (2015). Self-adaptive multiprecision preconditioners on multicore and Manycore architectures. In O. Marques, M. Dayde, & K. Nakajima (Eds.), High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers (pp. 115-123). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8969). Springer Verlag. https://doi.org/10.1007/978-3-319-17353-5_10

Anzt, Hartwig ; Lukarski, Dimitar ; Tomov, Stanimire et al. / Self-adaptive multiprecision preconditioners on multicore and Manycore architectures. High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers. editor / Osni Marques ; Michel Dayde ; Kengo Nakajima. Springer Verlag, 2015. pp. 115-123 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{4357494f68114f4591cda03b47b14765,

title = "Self-adaptive multiprecision preconditioners on multicore and Manycore architectures",

abstract = "Based on the premise that preconditioners needed for scientific computing are not only required to be robust in the numerical sense, but also scalable for up to thousands of light-weight cores, we argue that this two-fold goal is achieved for the recently developed self-adaptive multi-elimination preconditioner. For this purpose, we revise the underlying idea and analyze the performance of implementations realized in the PARALUTION and MAGMA open-source software libraries on GPU architectures (using either CUDA or OpenCL), Intel{\textquoteright}s Many Integrated Core Architecture, and Intel{\textquoteright}s Sandy Bridge processor. The comparison with other well-established preconditioners like multi-coloured Gauss- Seidel, ILU(0) and multi-colored ILU(0), shows that the twofold goal of a numerically stable cross-platform performant algorithm is achieved.",

author = "Hartwig Anzt and Dimitar Lukarski and Stanimire Tomov and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 11th International Conference on High Performance Computing for Computational Science, VECPAR 2014 ; Conference date: 30-06-2014 Through 03-07-2014",

year = "2015",

doi = "10.1007/978-3-319-17353-5\_10",

language = "English",

isbn = "9783319173528",

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

publisher = "Springer Verlag",

pages = "115--123",

editor = "Osni Marques and Michel Dayde and Kengo Nakajima",

booktitle = "High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers",

}

Anzt, H, Lukarski, D, Tomov, S & Dongarra, J 2015, Self-adaptive multiprecision preconditioners on multicore and Manycore architectures. in O Marques, M Dayde & K Nakajima (eds), High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8969, Springer Verlag, pp. 115-123, 11th International Conference on High Performance Computing for Computational Science, VECPAR 2014, Eugene, United States, 06/30/14. https://doi.org/10.1007/978-3-319-17353-5_10

Self-adaptive multiprecision preconditioners on multicore and Manycore architectures. / Anzt, Hartwig; Lukarski, Dimitar; Tomov, Stanimire et al.
High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers. ed. / Osni Marques; Michel Dayde; Kengo Nakajima. Springer Verlag, 2015. p. 115-123 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8969).

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

TY - GEN

T1 - Self-adaptive multiprecision preconditioners on multicore and Manycore architectures

AU - Anzt, Hartwig

AU - Lukarski, Dimitar

AU - Tomov, Stanimire

AU - Dongarra, Jack

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - Based on the premise that preconditioners needed for scientific computing are not only required to be robust in the numerical sense, but also scalable for up to thousands of light-weight cores, we argue that this two-fold goal is achieved for the recently developed self-adaptive multi-elimination preconditioner. For this purpose, we revise the underlying idea and analyze the performance of implementations realized in the PARALUTION and MAGMA open-source software libraries on GPU architectures (using either CUDA or OpenCL), Intel’s Many Integrated Core Architecture, and Intel’s Sandy Bridge processor. The comparison with other well-established preconditioners like multi-coloured Gauss- Seidel, ILU(0) and multi-colored ILU(0), shows that the twofold goal of a numerically stable cross-platform performant algorithm is achieved.

AB - Based on the premise that preconditioners needed for scientific computing are not only required to be robust in the numerical sense, but also scalable for up to thousands of light-weight cores, we argue that this two-fold goal is achieved for the recently developed self-adaptive multi-elimination preconditioner. For this purpose, we revise the underlying idea and analyze the performance of implementations realized in the PARALUTION and MAGMA open-source software libraries on GPU architectures (using either CUDA or OpenCL), Intel’s Many Integrated Core Architecture, and Intel’s Sandy Bridge processor. The comparison with other well-established preconditioners like multi-coloured Gauss- Seidel, ILU(0) and multi-colored ILU(0), shows that the twofold goal of a numerically stable cross-platform performant algorithm is achieved.

UR - https://www.scopus.com/pages/publications/84942626485

U2 - 10.1007/978-3-319-17353-5_10

DO - 10.1007/978-3-319-17353-5_10

M3 - Conference contribution

AN - SCOPUS:84942626485

SN - 9783319173528

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

SP - 115

EP - 123

BT - High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers

A2 - Marques, Osni

A2 - Dayde, Michel

A2 - Nakajima, Kengo

PB - Springer Verlag

T2 - 11th International Conference on High Performance Computing for Computational Science, VECPAR 2014

Y2 - 30 June 2014 through 3 July 2014

ER -

Anzt H, Lukarski D, Tomov S, Dongarra J. Self-adaptive multiprecision preconditioners on multicore and Manycore architectures. In Marques O, Dayde M, Nakajima K, editors, High Performance Computing for Computational Science - VECPAR 2014 - 11th International Conference, Revised Selected Papers. Springer Verlag. 2015. p. 115-123. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-17353-5_10

Self-adaptive multiprecision preconditioners on multicore and Manycore architectures

Abstract

Publication series

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this