Plasma: Parallel linear algebra software for multicore using OpenMP

Jack Dongarra; Mark Gates; Azzam Haidar; Jakub Kurzak; Piotr Luszczek; Panruo Wu; Ichitaro Yamazaki; Asim Yarkhan; Maksims Abalenkovs; Negin Bagherpour; Sven Hammarling; Jakub Šístek; David Stevens; Mawussi Zounon; Samuel D. Relton

doi:10.1145/3264491

Plasma: Parallel linear algebra software for multicore using OpenMP

Jack Dongarra, Mark Gates, Azzam Haidar, Jakub Kurzak, Piotr Luszczek, Panruo Wu, Ichitaro Yamazaki, Asim Yarkhan, Maksims Abalenkovs, Negin Bagherpour, Sven Hammarling, Jakub Šístek, David Stevens, Mawussi Zounon, Samuel D. Relton

Research output: Contribution to journal › Article › peer-review

36 Scopus citations

Abstract

The recent version of the Parallel Linear Algebra Software for Multicore Architectures (PLASMA) library is based on tasks with dependencies from the OpenMP standard. The main functionality of the library is presented. Extensive benchmarks are targeted on three recent multicore and manycore architectures, namely, an Intel Xeon, Intel Xeon Phi, and IBM POWER 8 processors.

Original language	English
Article number	a16
Journal	ACM Transactions on Mathematical Software
Volume	45
Issue number	2
DOIs	https://doi.org/10.1145/3264491
State	Published - Apr 2019
Externally published	Yes

Funding

The authors thank the Intel Corporation for their generous hardware donation and continuous financial support. They also thank the Oak Ridge Leadership Computing Facility for providing access to the POWER8 system. This work has been supported by the National Science Foundation under Grants No. CCF-1339822 (SILAS) and No. CCF-1527706 (DARE), by the European Commission H2020 Projects No. 671633 (NLAFET) and No. 671602 (INTERTWinE), and by the Engineering and Physical Sciences Research Council Project No. EP/M01147X/1 (SERT). Authors’ addresses: J. Dongarra, M. Gates, A. Haidar, J. Kurzak, P. Luszczek, I. Yamazaki, and A. YarKhan, University of Tennessee, Department of Electrical Engineering & Computer Science, 1122 Volunteer Blvd. Suite 203, Knoxville, TN 37996-3450, USA; emails: {dongarra, mgates3, haidar, kurzak, luszczek, iyamazak, yarkhan}@icl.utk.edu; P. Wu, University of Houston, Department of Computer Science, 3551 Cullen Blvd, Houston, TX 77204, USA; email: [email protected]; M. Abalenkovs, N. Bagherpour, S. Hammarling, and D. Stevens, The University of Manchester, School of Mathematics, Manchester, M13 9PL, UK; emails: [email protected], [email protected], sven. [email protected], [email protected]; J. Šístek, Institute of Mathematics of the Czech Academy of Sciences, Žitná 25, 115 67 Prague, Czech Republic; email: [email protected]; M. Zounon, Numerical Algorithms Group (NAG), Manchester One, 53 Portland Street Manchester, M1 3LD, UK; email: [email protected]; S. D. Relton, The University of Leeds, Institute of Health Sciences, Leeds, LS2 9LJ, UK; email: [email protected].

Funders	Funder number
SERT
SILAS	CCF-1527706
National Science Foundation	CCF-1339822
National Science Foundation
Intel Corporation
Horizon 2020 Framework Programme	671633
Horizon 2020 Framework Programme
Engineering and Physical Sciences Research Council	EP/M01147X/1
Engineering and Physical Sciences Research Council
European Commission	671602
European Commission

Keywords

Multicore processors
Numerical linear algebra libraries
OpenMP
PLASMA
Task-based programming
Tile algorithms

Access to Document

10.1145/3264491

Cite this

Dongarra, J., Gates, M., Haidar, A., Kurzak, J., Luszczek, P., Wu, P., Yamazaki, I., Yarkhan, A., Abalenkovs, M., Bagherpour, N., Hammarling, S., Šístek, J., Stevens, D., Zounon, M., & Relton, S. D. (2019). Plasma: Parallel linear algebra software for multicore using OpenMP. ACM Transactions on Mathematical Software, 45(2), Article a16. https://doi.org/10.1145/3264491

@article{78b6a0c6029743abaa968f396baeaa2f,

title = "Plasma: Parallel linear algebra software for multicore using OpenMP",

abstract = "The recent version of the Parallel Linear Algebra Software for Multicore Architectures (PLASMA) library is based on tasks with dependencies from the OpenMP standard. The main functionality of the library is presented. Extensive benchmarks are targeted on three recent multicore and manycore architectures, namely, an Intel Xeon, Intel Xeon Phi, and IBM POWER 8 processors.",

keywords = "Multicore processors, Numerical linear algebra libraries, OpenMP, PLASMA, Task-based programming, Tile algorithms",

author = "Jack Dongarra and Mark Gates and Azzam Haidar and Jakub Kurzak and Piotr Luszczek and Panruo Wu and Ichitaro Yamazaki and Asim Yarkhan and Maksims Abalenkovs and Negin Bagherpour and Sven Hammarling and Jakub {\v S}{\'i}stek and David Stevens and Mawussi Zounon and Relton, {Samuel D.}",

note = "Publisher Copyright: {\textcopyright} 2019 Copyright held by the owner/author(s).",

year = "2019",

month = apr,

doi = "10.1145/3264491",

language = "English",

volume = "45",

journal = "ACM Transactions on Mathematical Software",

issn = "0098-3500",

publisher = "Association for Computing Machinery (ACM)",

number = "2",

}

TY - JOUR

T1 - Plasma

T2 - Parallel linear algebra software for multicore using OpenMP

AU - Dongarra, Jack

AU - Gates, Mark

AU - Haidar, Azzam

AU - Kurzak, Jakub

AU - Luszczek, Piotr

AU - Wu, Panruo

AU - Yamazaki, Ichitaro

AU - Yarkhan, Asim

AU - Abalenkovs, Maksims

AU - Bagherpour, Negin

AU - Hammarling, Sven

AU - Šístek, Jakub

AU - Stevens, David

AU - Zounon, Mawussi

AU - Relton, Samuel D.

PY - 2019/4

Y1 - 2019/4

N2 - The recent version of the Parallel Linear Algebra Software for Multicore Architectures (PLASMA) library is based on tasks with dependencies from the OpenMP standard. The main functionality of the library is presented. Extensive benchmarks are targeted on three recent multicore and manycore architectures, namely, an Intel Xeon, Intel Xeon Phi, and IBM POWER 8 processors.

AB - The recent version of the Parallel Linear Algebra Software for Multicore Architectures (PLASMA) library is based on tasks with dependencies from the OpenMP standard. The main functionality of the library is presented. Extensive benchmarks are targeted on three recent multicore and manycore architectures, namely, an Intel Xeon, Intel Xeon Phi, and IBM POWER 8 processors.

KW - Multicore processors

KW - Numerical linear algebra libraries

KW - OpenMP

KW - PLASMA

KW - Task-based programming

KW - Tile algorithms

UR - http://www.scopus.com/inward/record.url?scp=85065709656&partnerID=8YFLogxK

U2 - 10.1145/3264491

DO - 10.1145/3264491

M3 - Article

AN - SCOPUS:85065709656

SN - 0098-3500

VL - 45

JO - ACM Transactions on Mathematical Software

JF - ACM Transactions on Mathematical Software

IS - 2

M1 - a16

ER -

Plasma: Parallel linear algebra software for multicore using OpenMP

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this