C++ Classes for Sparse Matrices and Iterative Linear Solvers

PDF

Authors
  1. Hally, D.
Corporate Authors
Defence Research Establishment Atlantic, Dartmouth NS (CAN)
Abstract
C++ classes for representing sparse matrices are described. The matrices use one of four profile storage schemes: compressed row storage, skyline storage, sparse skyline storage (similar to skyline storage but with compressed arrays above and below the diagonal), and S2dBlock storage (a special purpose storage scheme for matrices arising from partial differential equations on two-dimensional structured blocks). C++ classes are also provided for iterative solution methods for linear systems of equations. The Conjugate Gradient, Conjugate Gradient Squared, Quasi-Minimum Residual, and Generalized Minimum Residual (GMRES) methods have been implemented. Element re-ordering and preconditioning using the diagonal (Jacobi) or incomplete LU factorization of order zero (ILUO) methods are also provided.
Keywords
Sparse matrices
Report Number
DREA-TM-98-240 — Technical Memorandum
Date of publication
01 Oct 1998
Number of Pages
86
DSTKIM No
99-01768
CANDIS No
511873
Format(s):
Hardcopy;Document Image stored on Optical Disk

Permanent link

Document 1 of 1

Date modified: