Design, tuning and evaluation of parallel multilevel ILU preconditioners

José I. Aliaga; Matthias Bollhöfer; Alberto F. Martín; Enrique S. Quintana-Ortí

doi:10.1007/978-3-540-92859-1_28

Design, tuning and evaluation of parallel multilevel ILU preconditioners

José I. Aliaga, Matthias Bollhöfer, Alberto F. Martín, Enrique S. Quintana-Ortí

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

5 Citations (Scopus)

Abstract

In this paper, we present a parallel multilevel ILU preconditioner implemented with OpenMP. We employ METIS partitioning algorithms to decompose the computation into concurrent tasks, which are then scheduled to threads. Concretely, we combine decompositions which obtain significantly more tasks than processors, and the use of dynamic scheduling strategies in order to reduce the thread's idle time, which it is shown to be the main source of overhead in our parallel algorithm. Experimental results on a shared-memory platform consisting of 16 processors report remarkable performance for our approach.

Original language	English
Title of host publication	High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers
Pages	314-327
Number of pages	14
DOIs	https://doi.org/10.1007/978-3-540-92859-1_28
Publication status	Published - 2008
Externally published	Yes
Event	8th International Conference on High Performance Computing for Computational Science, VECPAR 2008 - Toulouse, France Duration: 24 Jun 2008 → 27 Jun 2008

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	5336 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	8th International Conference on High Performance Computing for Computational Science, VECPAR 2008
Country/Territory	France
City	Toulouse
Period	24/06/08 → 27/06/08

Access to Document

10.1007/978-3-540-92859-1_28

Cite this

Aliaga, J. I., Bollhöfer, M., Martín, A. F., & Quintana-Ortí, E. S. (2008). Design, tuning and evaluation of parallel multilevel ILU preconditioners. In High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers (pp. 314-327). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5336 LNCS). https://doi.org/10.1007/978-3-540-92859-1_28

Aliaga, José I. ; Bollhöfer, Matthias ; Martín, Alberto F. et al. / Design, tuning and evaluation of parallel multilevel ILU preconditioners. High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers. 2008. pp. 314-327 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3bdf4848c9d14cd59a1976480232f783,

title = "Design, tuning and evaluation of parallel multilevel ILU preconditioners",

abstract = "In this paper, we present a parallel multilevel ILU preconditioner implemented with OpenMP. We employ METIS partitioning algorithms to decompose the computation into concurrent tasks, which are then scheduled to threads. Concretely, we combine decompositions which obtain significantly more tasks than processors, and the use of dynamic scheduling strategies in order to reduce the thread's idle time, which it is shown to be the main source of overhead in our parallel algorithm. Experimental results on a shared-memory platform consisting of 16 processors report remarkable performance for our approach.",

keywords = "Incomplete LU factorization, OpenMP, Parallel algorithm, Shared-memory multiprocessor, Sparse linear system",

author = "Aliaga, {Jos{\'e} I.} and Matthias Bollh{\"o}fer and Mart{\'i}n, {Alberto F.} and Quintana-Ort{\'i}, {Enrique S.}",

year = "2008",

doi = "10.1007/978-3-540-92859-1_28",

language = "English",

isbn = "3540928588",

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

pages = "314--327",

booktitle = "High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers",

note = "8th International Conference on High Performance Computing for Computational Science, VECPAR 2008 ; Conference date: 24-06-2008 Through 27-06-2008",

}

Aliaga, JI, Bollhöfer, M, Martín, AF & Quintana-Ortí, ES 2008, Design, tuning and evaluation of parallel multilevel ILU preconditioners. in High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5336 LNCS, pp. 314-327, 8th International Conference on High Performance Computing for Computational Science, VECPAR 2008, Toulouse, France, 24/06/08. https://doi.org/10.1007/978-3-540-92859-1_28

Design, tuning and evaluation of parallel multilevel ILU preconditioners. / Aliaga, José I.; Bollhöfer, Matthias; Martín, Alberto F. et al.
High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers. 2008. p. 314-327 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5336 LNCS).

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

TY - GEN

T1 - Design, tuning and evaluation of parallel multilevel ILU preconditioners

AU - Aliaga, José I.

AU - Bollhöfer, Matthias

AU - Martín, Alberto F.

AU - Quintana-Ortí, Enrique S.

PY - 2008

Y1 - 2008

N2 - In this paper, we present a parallel multilevel ILU preconditioner implemented with OpenMP. We employ METIS partitioning algorithms to decompose the computation into concurrent tasks, which are then scheduled to threads. Concretely, we combine decompositions which obtain significantly more tasks than processors, and the use of dynamic scheduling strategies in order to reduce the thread's idle time, which it is shown to be the main source of overhead in our parallel algorithm. Experimental results on a shared-memory platform consisting of 16 processors report remarkable performance for our approach.

AB - In this paper, we present a parallel multilevel ILU preconditioner implemented with OpenMP. We employ METIS partitioning algorithms to decompose the computation into concurrent tasks, which are then scheduled to threads. Concretely, we combine decompositions which obtain significantly more tasks than processors, and the use of dynamic scheduling strategies in order to reduce the thread's idle time, which it is shown to be the main source of overhead in our parallel algorithm. Experimental results on a shared-memory platform consisting of 16 processors report remarkable performance for our approach.

KW - Incomplete LU factorization

KW - OpenMP

KW - Parallel algorithm

KW - Shared-memory multiprocessor

KW - Sparse linear system

UR - http://www.scopus.com/inward/record.url?scp=58349093301&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-92859-1_28

DO - 10.1007/978-3-540-92859-1_28

M3 - Conference contribution

SN - 3540928588

SN - 9783540928584

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

SP - 314

EP - 327

BT - High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers

T2 - 8th International Conference on High Performance Computing for Computational Science, VECPAR 2008

Y2 - 24 June 2008 through 27 June 2008

ER -

Aliaga JI, Bollhöfer M, Martín AF, Quintana-Ortí ES. Design, tuning and evaluation of parallel multilevel ILU preconditioners. In High Performance Computing for Computational Science - VECPAR 2008 - 8th International Conference, Revised Selected Papers. 2008. p. 314-327. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-540-92859-1_28

Design, tuning and evaluation of parallel multilevel ILU preconditioners

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this