[MITgcm-support] problem running tutorial_global_oce_latlon

[Samar Khatiwala]

Hi Jean-Michel

I use a home brewed opt file (very similar to the various *ia64_ifort*  
option files that ships with MITgcm).
See attached below.

I am able to run other configurations of the model compiled with this  
exact same opt file, so I'm puzzled
as to why it should fail on this particular one. Anyway, I'll try with  
a lower optimization level.

Thanks, Samar

FC='mpif77'
CC='mpicc'
LN='/bin/ln -s'
CPP='cpp  -traditional -P'
LINK=$FC

S64='$(TOOLSDIR)/set64bitConst.sh'

INCLUDES='-I/usr/local/mpich2/include'

DEFINES='-DALLOW_USE_MPI -DALWAYS_USE_MPI -DWORDLENGTH=4'

NOOPTFLAGS='-O0'
NOOPTFILES='mds_byteswapr8.F mds_byteswapr4.F mds_byteswapi4.F'

#  Note that the -mp switch is for ieee "maintain precision" and is
#  roughly equivalent to -ieee
if test "x$IEEE" = x ; then
     FOPTIM='-O3 -align'
     FFLAGS='-132 -r8 -i4 -w95 -W0 -WB -assume byterecl -convert  
big_endian'
else
     FOPTIM='-O0 -noalign'
     FFLAGS='-132 -r8 -i4 -w95 -W0 -WB -pc64 -xW -assume byterecl - 
convert big_endian'
fi


On Sep 15, 2008, at 4:12 PM, Jean-Michel Campin wrote:

> Hi samar,
>
> Which compiler options or optfile are you using ?
> Did you try with lower optimisation ?
>
> Jean-Michel
>
> On Mon, Sep 15, 2008 at 01:02:27PM -0400, Samar Khatiwala wrote:
>> Hi,
>>
>> I'm having problems with verification/tutorial_global_oce_latlon. The
>> model blows up (NaN's) within a
>> couple of time steps. I played around a bit with time steps (which
>> didn't help). Before I dig further I just
>> wanted to ask if anyone else has run into this problem.
>>
>> I am running this on a Linux Xeon box with ifort, gcc, and mpich2. I
>> don't have a problem running other
>> code.
>>
>> Thanks
>>
>> Samar
>>
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-support
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support