This is where to look when you are looking for source code in Fortran, C or
equivalent which implements numerical or statistical algorithms. Macros and
functions for common statistical computing environments are listed separately
under Statistical Computing Environments.
See also Free Statistical Programs.
- StatLib Index. A huge collection of statistical
software, data sets and links to statistics departments. The best place to start for
statisticians. Especially rich collection of Unix and S software. Mike Meyer,
- Allan Miller's Fortran
code. Least squares, subset selection in regression,
quadruple-precision, random number generation and optimization. Legacy code
written by Allan
- Index of Software. Department
of Statistics, Stanford University.
- Biomathematics Archive. Code for S,
LispStat, DOS and Macintosh. University of Texas MD Anderson Cancer Center.
- NetLib. The best collection of computational code
in one place. Maintained by AT&T Bell Laboratories, the University of Tennessee and
Oak Ridge National Laboratory.
- Guide to Available Mathematical Software (GAMS). A
heirarchical classification and virtual repository of mathematical and statistical
software, including packages from NetLib, TOMS, NAG and IMSL. Search for software by
problem or name. The GAMS database is restricted to software which meets recognized
standards of documentation and testing. National Institute of Standards and Technology.
- Collected Algorithms of the ACM.
Software associated with papers published in the Transactions on Mathematical Software
(TOMS) and other ACM journals. This software is refereed for originality, accuracy,
robustness, completeness, portability and lasting value. The collection is actually part
of NetLib, and is also indexed in GAMS. It is probably best searched via GAMS. Association
for Computing Machinery.
- GNU Scientific Library. For
any ANSI C compiler.
- Numerical Recipes. NetLib and GAMS can usually
provide code with more bells and whistles, but this is an excellent source when you want
code in Fortran or C which is understandable and compact. The entire text explaining the
methods as well as the code is available online. I wouldn't use Numerical Recipes for
production code, but the book is a beautifully readable introduction to many topics, and
the programs are useful to start you developing your own code. (It's worth noting that NR
has been criticized by numerical analysts; see the rebuttal
from NR.) Harvard University.
- Optimization Technology Center.
A joint enterprise of the Argonne National Laboratory and Northwestern University, with
the mission "to make potential users in industry, government, and academia aware of
how optimization techniques can aid their work, and to make the latest techniques widely
available". See the Optimization
- Numerical Methods. A
catalogue of sites dealing with numerical methods. By Tomasz Plewa.