Skip to content
mikiobraun edited this page Sep 13, 2010 · 2 revisions

Lapack is a pretty large library with hundreds of functions, but it basically covers the following four areas:

  • linear equations
  • linear least squares problems
  • eigenvalue problems
  • singular value problems

The large number of functions stems from the fact that these functions are implemented for up to four different data types (single, and double precision, real and complex numbers), and a number of different storage schemes (“normal” matrix, banded, tridiagonal, symmetric, etc.)

At the moment, jblas covers only a fraction of the possible methods, but I’m happy to add more functions if you need them. Just post a request on the jblas-users mailing list

Currently, the following methods are covered (only the “normal” matrix storage):

  • linear equations (Solve class)
    • general, symmetric and positive definite matrices.
  • linear least squares problems:
    • missing
  • eigenvalue problems (Eigen class)
    • general and symmetric matrices.
  • singular value problems
    • missing

In addition, jblas also covers some routines from the “computational routines” level:

  • matrix decomposition (Decompose class)
    • LU and Cholesky decomposition.