Impact of reordering on the memory of a multifrontal solver

被引：22

作者：

Guermouche, A ^{[1
]}

L'Excellent, JY ^{[1
]}

Utard, G ^{[1
]}

机构：

[1] Ecole Normale Super Lyon, CNRS, UMR 5668, INRIA,Lab Informat Parallelisme,ReMaP Project, F-69364 Lyon 07, France

来源：

PARALLEL COMPUTING | 2003年 / 29卷 / 09期

关键词：

sparse direct solvers; parallel multifrontal method; reordering; assembly tree; memory;

D O I：

10.1016/S0167-8191(03)00099-1

中图分类号：

TP301 [理论、方法];

学科分类号：

081202 ;

摘要：

This paper is concerned with the memory usage of sparse direct solvers, which depends on the ordering of the unknowns and the scheduling of the computational tasks. We study the influence of state-of-the-art sparse matrix reordering techniques on the memory usage of a multifrontal solver. Concerning the scheduling, the memory usage depends on the tree traversal and how the tasks are assigned to the processors. We analyze the memory scalability when a dynamic scheduling strategy mainly based on the balance of the workload is used. Finally we give hints to improve the parallel memory behaviour. (C) 2003 Elsevier B.V. All rights reserved.

引用

页码：1191 / 1218

页数：28

共 29 条

[1] An approximate minimum degree ordering algorithm [J].