Hide menu

6FMAI13 Linear Algebra Software packages

The computer exercises for the course are based on Matlab. The software package Matlab is primarily designed for performing numerical computations. Many of the standard mathematical operations are easily accessible within the Matlab environment.

A good free alternative to Matlab is SciLab. The SciLab environment is very similar to that of Matlab and in most cases programs developed for Matlab also works for SciLab with only minimal modifications.

For commercial or scientific applications where computationally demanding operations are carried out often a standard programming language (such as Fortran or C++) is used for efficiency reasons. For such applications scientific subroutine libraries are available for solving most common mathematical problems. Examples of freely available software packages include

  • Basic Linear Algebra Subprograms(BLAS). This package contains efficient implementations of basic linear algebra operations such as matrix-matrix multiplication.
  • LAPACK is a collection of Fortran routines for solving linear systems of equations, linear least squares problems, eigenvalue problems, computing the singular value decomposition, and more.
  • ARPACK is a Fortran package for solving large scale eigenvalue problems.
  • ATLAS or Automatically Tuned Linear Algebra Software contains efficient implementations of BLAS and most of LAPACK.
These and many more are available from www.netlib.org which acts as a repository for many scientific software packages.

Page responsible: Fredrik Berntsson
Last updated: 2020-06-24