Numerical linear algebra on emerging architectures: The PLASMA and MAGMA projects

Emmanuel Agullo; Jim Demmel; Jack Dongarra; Bilel Hadri; Jakub Kurzak; Julien Langou; Hatem Ltaief; Piotr Luszczek; Stanimire Tomov

doi:10.1088/1742-6596/180/1/012037

Numerical linear algebra on emerging architectures: The PLASMA and MAGMA projects

Emmanuel Agullo^*, Jim Demmel, Jack Dongarra, Bilel Hadri, Jakub Kurzak, Julien Langou, Hatem Ltaief, Piotr Luszczek, Stanimire Tomov

^*Corresponding author for this work

Research output: Contribution to journal › Conference article › peer-review

322 Scopus citations

Abstract

The emergence and continuing use of multi-core architectures and graphics processing units require changes in the existing software and sometimes even a redesign of the established algorithms in order to take advantage of now prevailing parallelism. Parallel Linear Algebra for Scalable Multi-core Architectures (PLASMA) and Matrix Algebra on GPU and Multics Architectures (MAGMA) are two projects that aims to achieve high performance and portability across a wide range of multi-core architectures and hybrid systems respectively. We present in this document a comparative study of PLASMA's performance against established linear algebra packages and some preliminary results of MAGMA on hybrid multi-core and GPU systems.

Original language	English (US)
Article number	012037
Journal	Journal of Physics: Conference Series
Volume	180
Issue number	1
DOIs	https://doi.org/10.1088/1742-6596/180/1/012037
State	Published - 2009
Externally published	Yes

ASJC Scopus subject areas

General Physics and Astronomy

Access to Document

10.1088/1742-6596/180/1/012037

Cite this

@article{835fdea45cdc4897bffee7a5da8c04be,

title = "Numerical linear algebra on emerging architectures: The PLASMA and MAGMA projects",

abstract = "The emergence and continuing use of multi-core architectures and graphics processing units require changes in the existing software and sometimes even a redesign of the established algorithms in order to take advantage of now prevailing parallelism. Parallel Linear Algebra for Scalable Multi-core Architectures (PLASMA) and Matrix Algebra on GPU and Multics Architectures (MAGMA) are two projects that aims to achieve high performance and portability across a wide range of multi-core architectures and hybrid systems respectively. We present in this document a comparative study of PLASMA's performance against established linear algebra packages and some preliminary results of MAGMA on hybrid multi-core and GPU systems.",

author = "Emmanuel Agullo and Jim Demmel and Jack Dongarra and Bilel Hadri and Jakub Kurzak and Julien Langou and Hatem Ltaief and Piotr Luszczek and Stanimire Tomov",

year = "2009",

doi = "10.1088/1742-6596/180/1/012037",

language = "English (US)",

volume = "180",

journal = "Journal of Physics: Conference Series",

issn = "1742-6588",

publisher = "IOP Publishing Ltd.",

number = "1",

}

TY - JOUR

T1 - Numerical linear algebra on emerging architectures

T2 - The PLASMA and MAGMA projects

AU - Agullo, Emmanuel

AU - Demmel, Jim

AU - Dongarra, Jack

AU - Hadri, Bilel

AU - Kurzak, Jakub

AU - Langou, Julien

AU - Ltaief, Hatem

AU - Luszczek, Piotr

AU - Tomov, Stanimire

PY - 2009

Y1 - 2009

N2 - The emergence and continuing use of multi-core architectures and graphics processing units require changes in the existing software and sometimes even a redesign of the established algorithms in order to take advantage of now prevailing parallelism. Parallel Linear Algebra for Scalable Multi-core Architectures (PLASMA) and Matrix Algebra on GPU and Multics Architectures (MAGMA) are two projects that aims to achieve high performance and portability across a wide range of multi-core architectures and hybrid systems respectively. We present in this document a comparative study of PLASMA's performance against established linear algebra packages and some preliminary results of MAGMA on hybrid multi-core and GPU systems.

AB - The emergence and continuing use of multi-core architectures and graphics processing units require changes in the existing software and sometimes even a redesign of the established algorithms in order to take advantage of now prevailing parallelism. Parallel Linear Algebra for Scalable Multi-core Architectures (PLASMA) and Matrix Algebra on GPU and Multics Architectures (MAGMA) are two projects that aims to achieve high performance and portability across a wide range of multi-core architectures and hybrid systems respectively. We present in this document a comparative study of PLASMA's performance against established linear algebra packages and some preliminary results of MAGMA on hybrid multi-core and GPU systems.

UR - http://www.scopus.com/inward/record.url?scp=77953997924&partnerID=8YFLogxK

U2 - 10.1088/1742-6596/180/1/012037

DO - 10.1088/1742-6596/180/1/012037

M3 - Conference article

AN - SCOPUS:77953997924

SN - 1742-6588

VL - 180

JO - Journal of Physics: Conference Series

JF - Journal of Physics: Conference Series

IS - 1

M1 - 012037

ER -

Numerical linear algebra on emerging architectures: The PLASMA and MAGMA projects

Abstract

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this