Software produced by members of the MaSe-team



A Superfast Solver for Real Symmetric Toeplitz Systems Using Real Trigonometric Transformations

Here the interested reader can download the software used to produce the experimental results in the report:


Semiseparable matrices and the symmetric eigenvalue problem

We refer the interested reader to the software corresponding to the PhD-thesis of Raf Vandebril.


An implicit QR-algorithm to compute the eigensystem of symmetric semiseparable matrices

The Matlab-files as a tarred-file.

Reference:


Solving diagonal-plus-semiseparable systems using a QR or a URV decomposition

The Matlab-files as a zipped-file.

Reference:


Reducing a symmetric matrix by orthogonal similarity transformations into a semiseparable matrix

and the link with the Lanczos-Ritz values

The Matlab-files as a zipped-file or as a tarred-file.

Reference


A fast (block) Hankel solver

Click here to download our fast Hankel solver.

The corresponding package for block Hankel matrices can be found here.

Both packages are written in Fortran 90.
They were compiled and tested using the xlf90 compiler on an IBM RS6000.

Matlab m-files are available via anonymous ftp to ftp.mathworks.com.
Look for the files hsolve.m, ratint.m, block_hsolve.m and block_ratint.m in the directory /pub/contrib/v4/linalg.

References:


A superfast Toeplitz solver

Click here to download our superfast Toeplitz solver.

This package is written in Fortran 90.
It was compiled and tested using the xlf90 compiler on an IBM RS6000 and the f90 compiler on a Sun SPARC.

Reference:


Solving a multivariate polynomial interpolation problem

Click here to download the implementation in Maple.

Reference: