Key concepts for parallel out-of-core LU factorization

被引：15

作者：

Dongarra, JJ ^{[1
]}

Hammarling, S

Walker, DW

机构：

[1] Univ Tennessee, Dept Comp Sci, Knoxville, TN 37996 USA

[2] NAG Ltd, Oxford OX2 8DR, England

[3] Oak Ridge Natl Lab, Math Sci Sect, Oak Ridge, TN 37831 USA

来源：

COMPUTERS & MATHEMATICS WITH APPLICATIONS | 1998年 / 35卷 / 07期

关键词：

out-of-core LU factorization; dense matrices; parallel computing; performance;

D O I：

10.1016/S0898-1221(98)00029-7

中图分类号：

O29 [应用数学];

学科分类号：

070104 ;

摘要：

This paper considers key ideas in the design of out-of-core dense LU factorization routines. A left-looking variant of the LU factorization algorithm is shown to require less I/O to disk than the right-looking variant, and is used to develop a parallel, out-of-core implementation. This implementation makes use of a small library of parallel I/O routines, together with ScaLAPACK and PBLAS routines. Results for runs on an Intel Paragon are presented and interpreted using a simple performance model.

引用

页码：13 / 31

页数：19

共 18 条

[1]

Anderson E., 1995, LAPACK USERS GUIDE

[2]

[Anonymous], 4 WORKSH INP OUTP PA

[3]

Choi J, 1992, P 4 S FRONT MASS PAR, P120

[4]

CHOI J, 1995, CS95283 U TENN DEP C

[5]

CHOI J, 1995, CS95292 U TENN DEP C

[6]

Corbett Peter, 1995, P WORKSHOP INPUTOUTP, P1

[7] IMPLEMENTING LINEAR ALGEBRA ALGORITHMS FOR DENSE MATRICES ON A VECTOR PIPELINE MACHINE [J].

DONGARRA, JJ ;

GUSTAVSON, FG ;

KARP, A .

SIAM REVIEW, 1984, 26 (01) :91-112

[8] AN EXTENDED SET OF FORTRAN BASIC LINEAR ALGEBRA SUBPROGRAMS [J].

DONGARRA, JJ ;

DUCROZ, J ;

HAMMARLING, S ;

HANSON, RJ .

ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 1988, 14 (01) :1-17

[9]

DONGARRA JJ, 1990, ACM T MATH SOFTWARE, V16, P1, DOI 10.1145/77626.79170

[10] LARGE DENSE NUMERICAL LINEAR ALGEBRA IN 1993 - THE PARALLEL COMPUTING INFLUENCE [J].

EDELMAN, A .

INTERNATIONAL JOURNAL OF SUPERCOMPUTER APPLICATIONS AND HIGH PERFORMANCE COMPUTING, 1993, 7 (02) :113-128

← 1 2 →