Вестник Нижегородского университета им. Н.И. Лобачевского. Серия: Социальные науки

16+

Title of Article

NOVEL LINEAR EQUATIONS SYSTEMS SOLVER FOR SPARSE SYMMETRIC POSITIVE-DEFINITE MATRIX

Issue	5
Date	2012

Article type	scientific article	UDC	519.612: 004.4
Pages	376-384	Keywords

Authors

Kozinov Evgeniy AleksandrovichNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo
Lebedev Ilya GennadevichNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo
Lebedev Sergey AleksandrovichNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo
Malova Anna YurevnaNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo
Meerov Iosif BorisovichNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo
Sysoev Aleksandr VladimirovichNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo
Filippenko Stanislav SergeevichNizhegorodskiy gosuniversitet im. N.I. Lobachevskogo

Abstract

A new solver is presented for sparse SLAE with a symmetric positive definite matrix. A phased scheme of sparse SLAE solution based on the Cholesky method is described. A supernodal approach has been used for the phased scheme implementation. Parallel computing for shared memory systems has been implemented and the experimental results are given and compared with the results of some other well-known solvers.

File (in Russian)