[MITgcm-support] Pb with compilation: make:

[Ed Hill]

On Wed, 2005-06-08 at 09:48 -0400, Nicolas Wienders wrote:
> Smiles, smiles, smiles, i could compile and run a few examples.
> Thank you so much for your help.

Hi Nicolas,

Excellent!  I'm a little disappointed that it took us so long to get
past these annoying build issues but very happy that you got it working!


> Attached is the option file i have used.

Thank you!  I'll check it into CVS.  Collecting optfiles that are known
to work is a very helpful way to save information about systems that we
(the MITgcm developers) can't log into to test.


> And by the way looking at the opt file, i just realized why we might
> had
> this problem: 
> 
> # **NOTES**
> #   o with Darwin you must remember to build on a ufs file system. The
> default Darwin file system (hfs) treats
> #     upper case and lower case file names as the same file. This
> breaks the current MITgcm build procedure.
> #   o tim.c should use uembellished cloc() are the procedure name for
> XLF
> #     ETIME (timers.F) and FDATE (utils.F) are not available with
> Darwin+XLF so need to be switched out.
> #
> 
> The same problem will be present when we upgrade to tiger i believe,
> unless we format the disk
> with the case sensitive option. I am correct?

Actually, those HFS notes are completely out-of-date and will disappear
when we update our web site this summer.  Many months ago we generalized
the build system so that it works just fine on HFS and within the Cygwin
(GNU on MS Windows) environment.

The problem that you were having was a name-mangling incompatibility
between the C and Fortran compiler that you were using.  Generally, the
name mangling differences can be fixed with the FC_NAMEMANGLE() macro
that we use.  But sometimes (particularly when we don't have access to a
machine) it can be rather frustrating to figure out the right macro.


> One more question, sorry to be a pain, do you think there is any hope
> that i might
> be able to compile using the GNU fortran compilers?

Yes!  We routinely use MITgcm with many versions of GCC.  We know that
it works well with the GCC v3.x compilers (using "g77") and we have had
some luck and some problems with the very new "gfortran" that ships with
GCC v4.x.  We expect that GCC v4.x will improve quickly since it is
quite new and actively developed.

So, we currently recommend GCC v3.x.

Ed

-- 
Edward H. Hill III, PhD
office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
             Cambridge, MA 02139-4307
emails:  eh3 at mit.edu                ed at eh3.com
URLs:    http://web.mit.edu/eh3/    http://eh3.com/
phone:   617-253-0098
fax:     617-253-4464