Christian Holm Christensen wrote:
> I'd like to direct your attention to the "GNU Scientific Library"
> (GSL) developed, among others, by people from Los Alamos. This is a
> very general library with many many features aimed for scientific
> use. It's written in ANSI C, and is very portable. Also, it's released
> under GPL, making it truely OpenSource. No wrappers in C++ exist, but
> as far as I can tell, it should be fairly easy to do. For a more
> complete description of the features in GSL, please look at
Hi,
I'd second this idea - at least as something to investigate. We've had
GSL available in PHENIX for about a year and it seems to work well. One
of the nice things about GSL is that it's being written by people with
knowledge of numerical computing; it's not trivial to write good
numerical software. I mean, I can write a series expansion for a Bessel
function as well as the next guy, but I'd have a hard time guaranteeing
the results are correct out to the least-significant digit.
(Prof. Kahan at Berkeley has some nice papers about the myriad ways in
which numerical implementations go awry:
http://www.cs.berkeley.edu/~wkahan/ )
GSL also has a extensive test suite to spot check the behavior of the
library.
It's also not quite true that there are _no_ C++ wrappers for GSL. The
Fermilab folks have wrappers for the special functions part of GSL:
http://www.fnal.gov/docs/working-groups/fpcltf/Pkg/SpecialFunctions/doc/SpecialFunctions.html
Whether the particular way they've chosen to do things is the best way
or not is a different question, but it's at least a place to start.
Cheers,
Dave
--
David Morrison Brookhaven National Laboratory phone: 631-344-5840
Physics Department, Bldg 510 C fax: 631-344-3253
Upton, NY 11973-5000 email: dave@bnl.gov
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:40 MET