Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators

Jakub Kurzak; Mark Gates; Ali Charara; Asim YarKhan; Ichitaro Yamazaki; Jack Dongarra

doi:10.1007/978-3-030-29400-7_35

Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators

Jakub Kurzak, Mark Gates, Ali Charara, Asim YarKhan, Ichitaro Yamazaki, Jack Dongarra

Computer Science and Mathematics Div

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

6 Scopus citations

Abstract

This work presents two implementations of linear solvers for distributed-memory machines with GPU accelerators—one based on the Cholesky factorization and one based on the LU factorization with partial pivoting. The routines are developed as part of the Software for Linear Algebra Targeting Exascale (SLATE) package, which represents a sharp departure from the traditional conventions established by legacy packages, such as LAPACK and ScaLAPACK. The article lays out the principles of the new approach, discusses the implementation details, and presents the performance results.

Original language	English
Title of host publication	Euro-Par 2019
Subtitle of host publication	Parallel Processing - 25th International Conference on Parallel and Distributed Computing, Proceedings
Editors	Ramin Yahyapour
Publisher	Springer
Pages	495-506
Number of pages	12
ISBN (Print)	9783030293994
DOIs	https://doi.org/10.1007/978-3-030-29400-7_35
State	Published - 2019
Event	25th International European Conference on Parallel and Distributed Computing, Euro-Par 2019 - Göttingen, Germany Duration: Aug 26 2019 → Aug 30 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11725 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th International European Conference on Parallel and Distributed Computing, Euro-Par 2019
Country/Territory	Germany
City	Göttingen
Period	08/26/19 → 08/30/19

Funding

This research was supported by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of two U.S. Department of Energy organizations (Office of Science and the National Nuclear Security Administration). Acknowledgments. This research was supported by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of two U.S. Department of Energy organizations (Office of Science and the National Nuclear Security Administration) responsible for the planning and preparation of a capable exascale ecosystem, including software, applications, hardware, advanced system engineering and early testbed platforms, in support of the nation’s exascale computing imperative.

Funders	Funder number
U.S. Department of Energy organizations
National Nuclear Security Administration
U.S. Department of Energy organizations
National Nuclear Security Administration

Keywords

Cholesky factorization
Distributed memory
GPU acceleration
LU factorization
Linear algebra
Linear systems of equations

Access to Document

10.1007/978-3-030-29400-7_35

Cite this

Kurzak, J., Gates, M., Charara, A., YarKhan, A., Yamazaki, I., & Dongarra, J. (2019). Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators. In R. Yahyapour (Ed.), Euro-Par 2019: Parallel Processing - 25th International Conference on Parallel and Distributed Computing, Proceedings (pp. 495-506). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11725 LNCS). Springer. https://doi.org/10.1007/978-3-030-29400-7_35

Kurzak, Jakub ; Gates, Mark ; Charara, Ali et al. / Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators. Euro-Par 2019: Parallel Processing - 25th International Conference on Parallel and Distributed Computing, Proceedings. editor / Ramin Yahyapour. Springer, 2019. pp. 495-506 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{e7e2ad19b2114c54a7608588af1fa8a2,

title = "Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators",

abstract = "This work presents two implementations of linear solvers for distributed-memory machines with GPU accelerators—one based on the Cholesky factorization and one based on the LU factorization with partial pivoting. The routines are developed as part of the Software for Linear Algebra Targeting Exascale (SLATE) package, which represents a sharp departure from the traditional conventions established by legacy packages, such as LAPACK and ScaLAPACK. The article lays out the principles of the new approach, discusses the implementation details, and presents the performance results.",

keywords = "Cholesky factorization, Distributed memory, GPU acceleration, LU factorization, Linear algebra, Linear systems of equations",

author = "Jakub Kurzak and Mark Gates and Ali Charara and Asim YarKhan and Ichitaro Yamazaki and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 25th International European Conference on Parallel and Distributed Computing, Euro-Par 2019 ; Conference date: 26-08-2019 Through 30-08-2019",

year = "2019",

doi = "10.1007/978-3-030-29400-7_35",

language = "English",

isbn = "9783030293994",

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

publisher = "Springer",

pages = "495--506",

editor = "Ramin Yahyapour",

booktitle = "Euro-Par 2019",

}

Kurzak, J, Gates, M, Charara, A, YarKhan, A, Yamazaki, I & Dongarra, J 2019, Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators. in R Yahyapour (ed.), Euro-Par 2019: Parallel Processing - 25th International Conference on Parallel and Distributed Computing, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11725 LNCS, Springer, pp. 495-506, 25th International European Conference on Parallel and Distributed Computing, Euro-Par 2019, Göttingen, Germany, 08/26/19. https://doi.org/10.1007/978-3-030-29400-7_35

Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators. / Kurzak, Jakub; Gates, Mark; Charara, Ali et al.
Euro-Par 2019: Parallel Processing - 25th International Conference on Parallel and Distributed Computing, Proceedings. ed. / Ramin Yahyapour. Springer, 2019. p. 495-506 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11725 LNCS).

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

TY - GEN

T1 - Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators

AU - Kurzak, Jakub

AU - Gates, Mark

AU - Charara, Ali

AU - YarKhan, Asim

AU - Yamazaki, Ichitaro

AU - Dongarra, Jack

PY - 2019

Y1 - 2019

N2 - This work presents two implementations of linear solvers for distributed-memory machines with GPU accelerators—one based on the Cholesky factorization and one based on the LU factorization with partial pivoting. The routines are developed as part of the Software for Linear Algebra Targeting Exascale (SLATE) package, which represents a sharp departure from the traditional conventions established by legacy packages, such as LAPACK and ScaLAPACK. The article lays out the principles of the new approach, discusses the implementation details, and presents the performance results.

AB - This work presents two implementations of linear solvers for distributed-memory machines with GPU accelerators—one based on the Cholesky factorization and one based on the LU factorization with partial pivoting. The routines are developed as part of the Software for Linear Algebra Targeting Exascale (SLATE) package, which represents a sharp departure from the traditional conventions established by legacy packages, such as LAPACK and ScaLAPACK. The article lays out the principles of the new approach, discusses the implementation details, and presents the performance results.

KW - Cholesky factorization

KW - Distributed memory

KW - GPU acceleration

KW - LU factorization

KW - Linear algebra

KW - Linear systems of equations

UR - http://www.scopus.com/inward/record.url?scp=85077117954&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-29400-7_35

DO - 10.1007/978-3-030-29400-7_35

M3 - Conference contribution

AN - SCOPUS:85077117954

SN - 9783030293994

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

SP - 495

EP - 506

BT - Euro-Par 2019

A2 - Yahyapour, Ramin

PB - Springer

T2 - 25th International European Conference on Parallel and Distributed Computing, Euro-Par 2019

Y2 - 26 August 2019 through 30 August 2019

ER -

Kurzak J, Gates M, Charara A, YarKhan A, Yamazaki I, Dongarra J. Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators. In Yahyapour R, editor, Euro-Par 2019: Parallel Processing - 25th International Conference on Parallel and Distributed Computing, Proceedings. Springer. 2019. p. 495-506. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-29400-7_35

Linear Systems Solvers for Distributed-Memory Machines with GPU Accelerators

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this