The parallelization of the incomplete LU factorization on AP1000

Takashi Nodera, Naoto Tsuno

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Using a finite difference method to discretize a two dimensional elliptic boundary value problem, we obtain systems of linear equations Ax = b, where the coefficient matrix A is a large, sparse, and nonsingular. These systems are often solved by preconditioned iterative methods. This paper presents a data distribution and a communication scheme for the parallelization of the preconditioner based on the incomplete LU factorization. At last, parallel performance tests of the preconditioner, using BiCGStab(ℓ) and GMRES (m) method, are carried out on a distributed memory parallel machine AP1000. The numerical results show that the preconditioner based on the incomplete LU factorization can be used even for MIMD parallel machines.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages788-792
Number of pages5
Volume1470 LNCS
Publication statusPublished - 1998
Event4th International Conference on Parallel Processing, Euro-Par 1998 - Southampton, United Kingdom
Duration: 1998 Sep 11998 Sep 4

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1470 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other4th International Conference on Parallel Processing, Euro-Par 1998
CountryUnited Kingdom
CitySouthampton
Period98/9/198/9/4

Fingerprint

Incomplete LU Factorization
Factorization
Parallelization
Preconditioner
Parallel Machines
Iterative methods
Linear equations
Finite difference method
Preconditioned Iterative Methods
Boundary value problems
GMRES
Performance Test
Elliptic Boundary Value Problems
Data Distribution
Distributed Memory
System of Linear Equations
Data storage equipment
Difference Method
Communication
Finite Difference

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Nodera, T., & Tsuno, N. (1998). The parallelization of the incomplete LU factorization on AP1000. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1470 LNCS, pp. 788-792). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1470 LNCS).

The parallelization of the incomplete LU factorization on AP1000. / Nodera, Takashi; Tsuno, Naoto.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1470 LNCS 1998. p. 788-792 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1470 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Nodera, T & Tsuno, N 1998, The parallelization of the incomplete LU factorization on AP1000. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 1470 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1470 LNCS, pp. 788-792, 4th International Conference on Parallel Processing, Euro-Par 1998, Southampton, United Kingdom, 98/9/1.
Nodera T, Tsuno N. The parallelization of the incomplete LU factorization on AP1000. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1470 LNCS. 1998. p. 788-792. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Nodera, Takashi ; Tsuno, Naoto. / The parallelization of the incomplete LU factorization on AP1000. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1470 LNCS 1998. pp. 788-792 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{3c8e6c12c0da4f1292ee7603ec1fe021,
title = "The parallelization of the incomplete LU factorization on AP1000",
abstract = "Using a finite difference method to discretize a two dimensional elliptic boundary value problem, we obtain systems of linear equations Ax = b, where the coefficient matrix A is a large, sparse, and nonsingular. These systems are often solved by preconditioned iterative methods. This paper presents a data distribution and a communication scheme for the parallelization of the preconditioner based on the incomplete LU factorization. At last, parallel performance tests of the preconditioner, using BiCGStab(ℓ) and GMRES (m) method, are carried out on a distributed memory parallel machine AP1000. The numerical results show that the preconditioner based on the incomplete LU factorization can be used even for MIMD parallel machines.",
author = "Takashi Nodera and Naoto Tsuno",
year = "1998",
language = "English",
isbn = "3540649522",
volume = "1470 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "788--792",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - The parallelization of the incomplete LU factorization on AP1000

AU - Nodera, Takashi

AU - Tsuno, Naoto

PY - 1998

Y1 - 1998

N2 - Using a finite difference method to discretize a two dimensional elliptic boundary value problem, we obtain systems of linear equations Ax = b, where the coefficient matrix A is a large, sparse, and nonsingular. These systems are often solved by preconditioned iterative methods. This paper presents a data distribution and a communication scheme for the parallelization of the preconditioner based on the incomplete LU factorization. At last, parallel performance tests of the preconditioner, using BiCGStab(ℓ) and GMRES (m) method, are carried out on a distributed memory parallel machine AP1000. The numerical results show that the preconditioner based on the incomplete LU factorization can be used even for MIMD parallel machines.

AB - Using a finite difference method to discretize a two dimensional elliptic boundary value problem, we obtain systems of linear equations Ax = b, where the coefficient matrix A is a large, sparse, and nonsingular. These systems are often solved by preconditioned iterative methods. This paper presents a data distribution and a communication scheme for the parallelization of the preconditioner based on the incomplete LU factorization. At last, parallel performance tests of the preconditioner, using BiCGStab(ℓ) and GMRES (m) method, are carried out on a distributed memory parallel machine AP1000. The numerical results show that the preconditioner based on the incomplete LU factorization can be used even for MIMD parallel machines.

UR - http://www.scopus.com/inward/record.url?scp=84882679877&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84882679877&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84882679877

SN - 3540649522

SN - 9783540649526

VL - 1470 LNCS

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

SP - 788

EP - 792

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

ER -