NMath Matrix provides a wide variety of structured sparse matrix types, including triangular, symmetric, Hermitian, banded, tridiagonal, symmetric banded, and Hermitian banded.
A sparse matrix is a matrix with only a small number of nonzero elements. A structured sparse matrix is one in which the zero elements (or elements contributing no new information) are distributed according to some pattern. By exploiting this pattern, structured sparse matrices can be manipulated more efficiently than general matrices, since all of the elements do not need to be stored.
This chapter describes the NMath Matrix structured sparse matrix types, and the storage schemes they use.
TOC | Previous | Next | Index