A column pre-ordering strategy for the unsymmetric-pattern multifrontal method

被引:254
作者
Davis, TA [1 ]
机构
[1] Univ Florida, Comp & Informat Sci & Engn Dept, Gainesville, FL 32611 USA
[2] Stanford Univ, Stanford, CA 94305 USA
[3] Univ Calif Berkeley, Lawrence Berkeley Lab, Berkeley, CA 94720 USA
来源
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE | 2004年 / 30卷 / 02期
关键词
algorithms; experimentation; performance; sparse nonsymmetric matrices; linear equations; multifrontal method; ordering methods;
D O I
10.1145/992200.992205
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
A new method for sparse LU factorization is presented that combines a column pre-ordering strategy with a right-looking unsymmetric-pattern multifrontal numerical factorization. The column ordering is selected to give a good a priori upper bound on fill-in and then refined during numerical factorization ( while preserving the bound). Pivot rows are selected to maintain numerical stability and to preserve sparsity. The method analyzes the matrix and automatically selects one of three pre-ordering and pivoting strategies. The number of nonzeros in the LU factors computed by the method is typically less than or equal to those found by a wide range of unsymmetric sparse LU factorization methods, including left-looking methods and prior multifrontal methods.
引用
收藏
页码:165 / 195
页数:31
相关论文
共 58 条
[1]   An approximate minimum degree ordering algorithm [J].
Amestoy, PR ;
Davis, TA ;
Duff, IS .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 1996, 17 (04) :886-905
[2]   An unsymmetrized multifrontal LU factorization [J].
Amestoy, PR ;
Puglisi, C .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2002, 24 (02) :553-569
[3]   Analysis and comparison of two general sparse solvers for distributed memory computers [J].
Amestoy, PR ;
Duff, IS ;
L'Excellent, JY ;
Li, XS .
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 2001, 27 (04) :388-421
[4]   VECTORIZATION OF A MULTIPROCESSOR MULTIFRONTAL CODE [J].
AMESTOY, PR ;
DUFF, IS .
INTERNATIONAL JOURNAL OF SUPERCOMPUTER APPLICATIONS AND HIGH PERFORMANCE COMPUTING, 1989, 3 (03) :41-59
[5]   A fully asynchronous multifrontal solver using distributed dynamic scheduling [J].
Amestoy, PR ;
Duff, IS ;
L'Excellent, JY ;
Koster, J .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2001, 23 (01) :15-41
[6]  
Amestoy PR, 1996, NUMER LINEAR ALGEBR, V3, P275
[7]  
AMESTOY PR, 1996, 2 SIAM C SPARS MATR
[8]  
AMESTOY PR, 2004, IN PRESS ACM T MATH
[9]   SOLVING SPARSE LINEAR-SYSTEMS WITH SPARSE BACKWARD ERROR [J].
ARIOLI, M ;
DEMMEL, JW ;
DUFF, IS .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 1989, 10 (02) :165-190
[10]   Robust ordering of sparse matrices using multisection [J].
Ashcraft, C ;
Liu, JWH .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 1998, 19 (03) :816-832