Non-JPL Sources of Mathematical Software

Unfortunately, while I wasn't looking, many of these links went 404. I've changed them to strikeout font.

There are several archives of mathematical software. A few are:

FFT has been done several times. Here are a few references

Greg Henry et. al. have a library of FFT and other routines for Linux on Intel Pentium.

Ernst Meese of the Norwegian University of Science and Technology, Trondheim, Norway has written a small library in Fortran 90 for solving sparse matrix problems. It contains modules for the compressed sparse row (CSR) and the modified sparse row (MSR) data storeage formats, a very general incomplete LU factorsation routine, the Krylov subspace solvers CGS, BiCGSTAB and GMRES, and some more. understandable. It was available together with a manual by ftp at ftp.maskin.ntnu.no/pub/mtf/smlib.
Unfortunately, that FTP server no longer accepts anonymous logins, so I have put my copy, which I got in September 1998, in a publicly accessible place.

Alan Miller has an interesting collection of links to mathematical and other software in Fortran 90.

Tomas Skalicky has developed a system for Large Sparse Linear Systems, called LASPack.

Henry Wolkowicz at Waterloo University has offered us access to his library of Combinatorial Optimization software.

Some computations require calculation of both function values and derivatives. It is at best tedious, and frequently quite difficult, to produce a correct program that computes derivatives. To aid in this problem, there is a program ADIFOR, which means Automatic Differentiation of Fortran programs. It accepts a Fortran subprogram to compute a function, and produces a Fortran subprogram that computes the derivative (Jacobian) of the function. There is also a collection of software for automatic differentiation (AD) that provides short highlights of Fortran 77, Fortran 90, ANSI-C and C++ AD tools, as well as of systems integrating AD such as MAPLE or AMPL, and AD support libraries.

There are several lists of Mathematics resources in the Web:

Commercial providers include:

We recommend you do not use Numerical Recipes.