The Bluebit.MatrixLibrary namespace contains the Matrix and Vector classes for representing real matrices and vectors and their complex equivalents CMatrix and CVector for complex matrices and vectors.

Additionally, it contains the Cholesky, LQ, LU, QR, SVD classes for factoring real matrices and the CCholesky, CLQ, CLU, CQR,  CSVD classes for complex matrices respectively.

The Eigen and CEigen classes are used for the eigenvalues and eigenvectors calculations of general matrices, at the same time as SymEigen for symmetric and CSymEigen for hermitian matrices.

Starting from version 5.0 Bluebit.MatrixLibrary offers support for sparse matrices.

  The SparseMatrix and CSparseMatrix classes are used to represent sparse real and complex matrices respectively. The constructor of the SparseMatrix allows to define a matrix as symmetric making possible to store only the upper diagonal elements. This reduces the required storage and allows faster computations. In the same way the constructor of the CSparseMatrix class allows to define a complex matrix as a hermitian matrix.

 The SparseSolver and CSparseSolver Class classes are used to solve effectively large sparse systems of linear equations. They allow to reuse the underlying LU factorization of the matrix for different right hands but as well they allow the reuse of the factorization when the underlying matrix is modified. In this later case an attempt is made to use a pre-existing factorization, compute an inexact solution at first and then improve the solution by using iterative methods. In case that significant changes have been made to the underlying matrix and the iterative methods fail, then a new factorization is computed automatically.