Automated empirical optimizations of software and the ATLAS project

R. Clint Whaley; Antoine Petitet; Jack J. Dongarra

doi:10.1016/S0167-8191(00)00087-9

Automated empirical optimizations of software and the ATLAS project

R. Clint Whaley, Antoine Petitet, Jack J. Dongarra

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

939 Scopus citations

Abstract

This paper describes the automatically tuned linear algebra software (ATLAS) project, as well as the fundamental principles that underly it. ATLAS is an instantiation of a new paradigm in high performance library production and maintenance, which we term automated empirical optimization of software (AEOS); this style of library management has been created in order to allow software to keep pace with the incredible rate of hardware advancement inherent in Moore's Law. ATLAS is the application of this new paradigm to linear algebra software, with the present emphasis on the basic linear algebra subprograms (BLAS), a widely used, performance-critical, linear algebra kernel library.

Original language	English
Pages (from-to)	3-35
Number of pages	33
Journal	Parallel Computing
Volume	27
Issue number	1-2
DOIs	https://doi.org/10.1016/S0167-8191(00)00087-9
State	Published - Jan 2001
Externally published	Yes

Funding

This work was supported in part by US Department of Energy under contract number DE-AC05-96OR22464; National Science Foundation Science and Technology Center Cooperative Agreement No. CCR-8809615; Los Alamos National Laboratory, University of California, subcontract #B76680017-3Z.

Funders	Funder number
U.S. Department of Energy	DE-AC05-96OR22464
University of California	76680017-3Z
Los Alamos National Laboratory
BioXFEL Science and Technology Center	CCR-8809615

Access to Document

10.1016/S0167-8191(00)00087-9

Cite this

@article{e8a7aea4037a43b2aa420b3f2b595aef,

title = "Automated empirical optimizations of software and the ATLAS project",

abstract = "This paper describes the automatically tuned linear algebra software (ATLAS) project, as well as the fundamental principles that underly it. ATLAS is an instantiation of a new paradigm in high performance library production and maintenance, which we term automated empirical optimization of software (AEOS); this style of library management has been created in order to allow software to keep pace with the incredible rate of hardware advancement inherent in Moore's Law. ATLAS is the application of this new paradigm to linear algebra software, with the present emphasis on the basic linear algebra subprograms (BLAS), a widely used, performance-critical, linear algebra kernel library.",

author = "{Clint Whaley}, R. and Antoine Petitet and Dongarra, {Jack J.}",

year = "2001",

month = jan,

doi = "10.1016/S0167-8191(00)00087-9",

language = "English",

volume = "27",

pages = "3--35",

journal = "Parallel Computing",

issn = "0167-8191",

publisher = "Elsevier",

number = "1-2",

}

TY - JOUR

T1 - Automated empirical optimizations of software and the ATLAS project

AU - Clint Whaley, R.

AU - Petitet, Antoine

AU - Dongarra, Jack J.

PY - 2001/1

Y1 - 2001/1

N2 - This paper describes the automatically tuned linear algebra software (ATLAS) project, as well as the fundamental principles that underly it. ATLAS is an instantiation of a new paradigm in high performance library production and maintenance, which we term automated empirical optimization of software (AEOS); this style of library management has been created in order to allow software to keep pace with the incredible rate of hardware advancement inherent in Moore's Law. ATLAS is the application of this new paradigm to linear algebra software, with the present emphasis on the basic linear algebra subprograms (BLAS), a widely used, performance-critical, linear algebra kernel library.

AB - This paper describes the automatically tuned linear algebra software (ATLAS) project, as well as the fundamental principles that underly it. ATLAS is an instantiation of a new paradigm in high performance library production and maintenance, which we term automated empirical optimization of software (AEOS); this style of library management has been created in order to allow software to keep pace with the incredible rate of hardware advancement inherent in Moore's Law. ATLAS is the application of this new paradigm to linear algebra software, with the present emphasis on the basic linear algebra subprograms (BLAS), a widely used, performance-critical, linear algebra kernel library.

UR - http://www.scopus.com/inward/record.url?scp=0343462141&partnerID=8YFLogxK

U2 - 10.1016/S0167-8191(00)00087-9

DO - 10.1016/S0167-8191(00)00087-9

M3 - Article

AN - SCOPUS:0343462141

SN - 0167-8191

VL - 27

SP - 3

EP - 35

JO - Parallel Computing

JF - Parallel Computing

IS - 1-2

ER -

Automated empirical optimizations of software and the ATLAS project

Abstract

Funding

Access to Document

Other files and links

Fingerprint

Cite this