A Parallel Algebraic Multigrid Solver on Graphics Processing Units

Gundolf Haase; Manfred Liebmann; Craig C. Douglas; Gernot Plank

doi:10.1007/978-3-642-11842-5_5

A Parallel Algebraic Multigrid Solver on Graphics Processing Units

Gundolf Haase, Manfred Liebmann, Craig C. Douglas, Gernot Plank

Research output: Chapter in Book/Report/Conference proceeding › Chapter

54 Scopus citations

Abstract

The paper presents a multi-GPU implementation of the preconditioned conjugate gradient algorithm with an algebraic multigrid preconditioner (PCG-AMG) for an elliptic model problem on a 3D unstructured grid. An efficient parallel sparse matrix-vector multiplication scheme underlying the PCG-AMG algorithm is presented for the many-core GPU architecture. A performance comparison of the parallel solver shows that a singe Nvidia Tesla C1060 GPU board delivers the performance of a sixteen node Infiniband cluster and a multi-GPU configuration with eight GPUs is about 100 times faster than a typical server CPU core. © 2010 Springer-Verlag.

Original language	English (US)
Title of host publication	High Performance Computing and Applications
Publisher	Springer Nature
Pages	38-47
Number of pages	10
ISBN (Print)	9783642118418
DOIs	https://doi.org/10.1007/978-3-642-11842-5_5
State	Published - 2010
Externally published	Yes

Access to Document

10.1007/978-3-642-11842-5_5

Cite this

@inbook{561285fd92554943ae384904d5ea9a39,

title = "A Parallel Algebraic Multigrid Solver on Graphics Processing Units",

abstract = "The paper presents a multi-GPU implementation of the preconditioned conjugate gradient algorithm with an algebraic multigrid preconditioner (PCG-AMG) for an elliptic model problem on a 3D unstructured grid. An efficient parallel sparse matrix-vector multiplication scheme underlying the PCG-AMG algorithm is presented for the many-core GPU architecture. A performance comparison of the parallel solver shows that a singe Nvidia Tesla C1060 GPU board delivers the performance of a sixteen node Infiniband cluster and a multi-GPU configuration with eight GPUs is about 100 times faster than a typical server CPU core. {\textcopyright} 2010 Springer-Verlag.",

author = "Gundolf Haase and Manfred Liebmann and Douglas, {Craig C.} and Gernot Plank",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledged KAUST grant number(s): KUS-C1-016-04 Acknowledgements: This publication is based on work supported in part by NSF grants OISE-0405349, ACI-0305466, CNS-0719626, and ACI-0324876, by DOE project DE-FC26-08NT4, by FWF project SFB032, by BMWF project AustrianGrid 2, and Award No. KUS-C1-016-04, made by King Abdullah University of Science and Technology (KAUST). This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2010",

doi = "10.1007/978-3-642-11842-5_5",

language = "English (US)",

isbn = "9783642118418",

pages = "38--47",

booktitle = "High Performance Computing and Applications",

publisher = "Springer Nature",

}

TY - CHAP

T1 - A Parallel Algebraic Multigrid Solver on Graphics Processing Units

AU - Haase, Gundolf

AU - Liebmann, Manfred

AU - Douglas, Craig C.

AU - Plank, Gernot

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledged KAUST grant number(s): KUS-C1-016-04 Acknowledgements: This publication is based on work supported in part by NSF grants OISE-0405349, ACI-0305466, CNS-0719626, and ACI-0324876, by DOE project DE-FC26-08NT4, by FWF project SFB032, by BMWF project AustrianGrid 2, and Award No. KUS-C1-016-04, made by King Abdullah University of Science and Technology (KAUST). This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2010

Y1 - 2010

N2 - The paper presents a multi-GPU implementation of the preconditioned conjugate gradient algorithm with an algebraic multigrid preconditioner (PCG-AMG) for an elliptic model problem on a 3D unstructured grid. An efficient parallel sparse matrix-vector multiplication scheme underlying the PCG-AMG algorithm is presented for the many-core GPU architecture. A performance comparison of the parallel solver shows that a singe Nvidia Tesla C1060 GPU board delivers the performance of a sixteen node Infiniband cluster and a multi-GPU configuration with eight GPUs is about 100 times faster than a typical server CPU core. © 2010 Springer-Verlag.

AB - The paper presents a multi-GPU implementation of the preconditioned conjugate gradient algorithm with an algebraic multigrid preconditioner (PCG-AMG) for an elliptic model problem on a 3D unstructured grid. An efficient parallel sparse matrix-vector multiplication scheme underlying the PCG-AMG algorithm is presented for the many-core GPU architecture. A performance comparison of the parallel solver shows that a singe Nvidia Tesla C1060 GPU board delivers the performance of a sixteen node Infiniband cluster and a multi-GPU configuration with eight GPUs is about 100 times faster than a typical server CPU core. © 2010 Springer-Verlag.

UR - http://hdl.handle.net/10754/597368

UR - http://link.springer.com/10.1007/978-3-642-11842-5_5

UR - http://www.scopus.com/inward/record.url?scp=77951541240&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-11842-5_5

DO - 10.1007/978-3-642-11842-5_5

M3 - Chapter

SN - 9783642118418

SP - 38

EP - 47

BT - High Performance Computing and Applications

PB - Springer Nature

ER -

A Parallel Algebraic Multigrid Solver on Graphics Processing Units

Abstract

Access to Document

Other files and links

Fingerprint

Cite this