ACADEMIA

Newly created example programs for LAPACK, (a library of numerical linear algebra subroutines designed for high performance on workstations, vector computers and shared memory processors), have been released by the Numerical Algorithms Group (NAG). These example programs are designed to help less experienced users to call LAPACK routines correctly. NAG, arguably the world’s leading provider of mathematical and statistical software for developers, has been involved in the community effort to create LAPACK since the inception of the LAPACK project. Through his active contribution to the LAPACK project, NAG Principal Consultant, Sven Hammarling has ensured that the entire LAPACK 3 routines and corrections have been included in NAG’s Library software. Hammarling explains that the example programs were created as a service to the open source community. He comments, “From monitoring the LAPACK list and forum I could see that example programs would be very useful to help many users get started with LAPACK, and would probably help many existing users as well. The example programs have been extracted from NAG Library example programs and made LAPACK-friendly. To the best of my knowledge, these are the first and only such LAPACK example programs that have been developed for the open source community.” LAPACK (Linear Algebra PACKage) provides routines for solving systems of simultaneous linear equations, linear least-squares problems, eigenvalue problems, and singular value problems. Routines for the associated matrix factorizations are also provided, as are related computations such as reordering Schur factorizations, condition number estimation and forward and backward error estimation. LAPACK is intended for dense and banded matrices, but not general sparse matrices. The original goal of the LAPACK project was to make the widely used EISPACK and LINPACK libraries run efficiently on shared-memory vector and parallel processors. The new LAPACK example programs provide an example of the use of each of the double precision LAPACK drivers. Each example comprises a description of the problem, together with links to the example program, the example data and example result. Non-LAPACK routines, mainly matrix printing routines, are provided as free downloads by NAG.