This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: gfortran on macs
- From: Brian Barnes <bcbarnes at artsci dot wustl dot edu>
- To: Brooks Moses <bmoses at stanford dot edu>
- Cc: fortran at gcc dot gnu dot org
- Date: Wed, 25 Jul 2007 12:30:59 -0500 (CDT)
- Subject: Re: gfortran on macs
- References: <4.3.1.2.20070723145324.026c0198@cits1.stanford.edu>
I believe that using the GFortran 4.3 binary package should work fine with a
system GCC of version 4.0 -- the GFortran package should be set up so as not
to interfere with the system library at all. But that's an educated guess; I
Thanks for the reply, Brooks. I tried the gfortran 4.3 binary package on
the official wiki / website. Unfortunately, it failed to complete the
install. I noticed that it installed many (possibly all) of its files in
/usr/local/gfortran as expected, but failed to create /usr/bin/gfortran.
There may have been other problems with the install, but I didn't see a
log or error message. This is on OS X 10.4.10 with the XCode from the
10.4.9 DVDs installed. All new stuff. The package was run with su
permissions. So I deleted what the gfortran pkg installed and went to
plan B...
I found it almost trivially simple to compile things from source, much like
it would be on a Linux machine (and the commands are identical, too.) The
I ended up fixing a problem with my fink install so that I could see the
gcc42 package in fink -stable and -unstable. Thank you Jack Howarth for
maintaining this package! I used 'sudo fink install gcc42' and its
install script detected the gmp and mpfr dependencies, and everything was
compiled from source and installed fine. I now have gcc 4.2.1 installed.
Building these for Intel OS X is supposedly non-trivial; gmp specifies
that they do not support Intel OS X in their 4.x releases, and building
of gcc required some multilib changes. There are other Intel OS X patches
laying around to fix build bugs too. But, I think J.H. took care of all
these things in his fink build script.
I tested the gfotran 4.2.1 install using our group's molecular dynamics
code (which is mostly F77, but is a numerically strong test) and
everything works. :)
Yeah, so I know this is a bit of a tangent from the normal development
talk here, but I hope some people find it helpful. Thank you, and good
luck with the work on 4.3!
Sincerely,
Brian Barnes