Sparse matrices

A sparse matrix is a matrix which number of non-zero terms is very small compared to its size. It presents some great advantages, like reducing the memory space required to store it and reducing the global number of arithmetic operations used for computations of the product matrix-vector for instance. Nowdays, many numerical and efficient algorithms are based on sparse linear algebra (PDE, Graph theory,..., see [Saad96]). In such applications it is not rare to treat (sparse) matrices of size or even , which could not be managed as dense matrices.

It is implemented as

template<class C, class R=sparse::rep2d<C> > struct MatrSps;

where

• C is the type of coefficients.

• R is the type of the internal representation.

Container

• The default value for the container is sparse::rep2d.

Implementation

See also:

synaps/linalg/MatrSps.h.