Article information

1998 , Volume 3, ¹ 1, p.23-30

Balandin M.Y., Shurina E.P.

Some Estimations of Efficiency for parallel SLE solving algorithms of Krylov Sequence Type

The ABR1ORT method for solving densed ill-conditioned systems of linear equations (SLEs) and the GMRES method for solving sparsed non-symmetric SLEs are discussed; both them belong to Krylov Sequence Methods. The analysis ofnumerical complexity and memory requirements is performed for these methods; parallel algorythms for multiprocessor computers are briefly discussed. For these algorythms, asymptotic estimations of parallel speedup are shown.

[full text] Classificator Msc2000:
*65F10 Iterative methods for linear systems
65F50 Sparse matrices
65Y05 Parallel computation
65Y20 Complexity and performance of numerical algorithms
Classificator Computer Science:
*F.2.1 Numerical Algorithms and Problems
G.4 Mathematical Software
G.1.0 General (Numerical Analysis)
G.1.3 Numerical Linear Algebra

Keywords: large-scale non-symmetric linear systems with sparse matrices, shared-memory multiprocessor, parallel computations, Krylov subspace method, ABR1ORT method, dense ill-conditioned systems, GMRES method, numerical complexity, numerical experiments

Author(s):
Balandin M.Yu.
Office: Novosibirsk State Technical University
Address: 630092, Russia, Novosibirsk, Marx ave. 20

Shurina Ella Petrovna
Dr. , Professor
Position: General Scientist
Office: Novosibirsk State Technical University, Trofimuk Institute of Petroleum Geology and Geophysics SB RAS
Address: 630073, Russia, Novosibirsk, Karl Marx Ave. 20
Phone Office: (343) 223-72-95
E-mail: shurina@online.sinor.ru


Bibliography link:
Balandin M.Y., Shurina E.P. Some Estimations of Efficiency for parallel SLE solving algorithms of Krylov Sequence Type // Computational technologies. 1998. V. 3. ¹ 1. P. 23-30
Home| Scope| Editorial Board| Content| Search| Subscription| Rules| Contacts
ISSN 1560-7534
© 2024 FRC ICT