Prev: INTRINSIC STATEMENT for functions overloaded for user defined types
Next: Passing optional parameters
From: analyst41 on 19 Sep 2009 11:26
On Sep 19, 9:56 am, Gus Gassmann <horand.gassm...(a)gmail.com> wrote:
> analys...(a)hotmail.com wrote:
> > On Sep 17, 12:16 pm, Bart Vandewoestyne
> > <MyFirstName.MyLastN...(a)telenet.be> wrote:
> >> Hello,
>
> >> I recently got asked the question again whether the NAG Fortran
> >> Library was available on our server.  In the past, I already got
> >> the same question about the IMSL library...
>
> >> There is one problem for me to answer that question: I like Open
> >> Source Software...
>
> >> So I would like to advise the persons asking me those questions
> >> with an 'Open Source answer'.
>
> >> To be honest... i always stick to the Fortran 95 standard when
> >> programming, and up until now i could implement all the
> >> 'numerical code' that i needed myself, without the use of
> >> external libraries but for example using Numerical Recipes or
> >> other descriptions of algorithms in papers.
>
> >> So... I have no experience with for example the GNU Scientific
> >> Library in combination with Fortran 95 (or 2003).  I know there
> >> exists a Fortran interface [1], but i have not tested it.
>
> >> So my questions to this group are:
>
> >> * Next to GSL, Netlib, GAMS and Numerical Recipes, what other
> >>   Open Source libraries are there that can easily be used
> >>   from a Fortran 95 program and are a valuable alternative to the
> >>   NAG Fortran Library or the IMSL?
>
> >> * Does anyone have experience with FGSL and would you advise its
> >>   use?
>
> >> Feel free to share your ideas/comments.  I want Open Source to
> >> rule at my institution! ;-)
>
> >> Kind regards,
> >> Bart
>
> >> [1]http://www.lrz-muenchen.de/services/software/mathematik/gsl/fortran/i...
>
> >> --
> >>         "Share what you know.  Learn what you don't."
>
> > There is COIN-OR for constrained optimization.  It was sponsored by
> > IBM but now they have bought a commercial vendor (ILOG) - and so its
> > not clear what level of IBM support one can expect going forward.
>
> >http://www.coin-or.org/
>
> COIN-OR was spun out of IBM several years ago and now exists as an
> independent entity. Development on key COIN-OR projects is continuing
> daily. On the other hand, almost all the development is done in C++.- Hide quoted text -
>
> - Show quoted text -

Thats good to know. Even if they don;t have fortran callable
libraries, one can always build problems in a standard format with
Fortran and use their routines as a black box to get the solution and
process the output with the language of one's choice.
First  |  Prev  | 
Pages: 1 2
Prev: INTRINSIC STATEMENT for functions overloaded for user defined types
Next: Passing optional parameters