This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: Incomplete Fermi Dirac ?
- From: Peter Teuben <teuben at astro dot umd dot edu>
- To: giorgio mugnaini <o1623499 at iet dot unipi dot it>
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Tue, 13 Apr 2004 11:49:41 -0400 (EDT)
- Subject: Re: Incomplete Fermi Dirac ?
- References: <407C0092.5060400@iet.unipi.it>
i have some ideas why it fails in gcc, i'v e often run into this
(g77 is much more ANSI strict often than most compilers), but
i did n't look into the code itself. Try this:
g77 -fzeros -finit-local-zero -fno-automatic -g -o icd icd.f
and i got better results. At least comparing the intel and gnu
compiler.
- peter
On Tue, 13 Apr 2004, giorgio mugnaini wrote:
> Hello,
> I'm with University of Pisa and I've found that algorithm ACM 745
> computes incomplete Fermi-Dirac integrals.
>
> An implementation is available in fortran 77 at:
>
> http://www.netlib.org/toms/745
>
> I've thought that this implementation could be used in order to build a
> port in C that can be included in GSL.
>
> The program works fine with Lahey Fortran and Portland Group Fortran,
> but gives wrong results with G77 (bugs?).
> Curiously, if it is translated in C with f2c, it works fine also with GCC!
>