[MITgcm-support] Limitation for Tile's dimensions
Eli Biton
eliyahub at wisemail.weizmann.ac.il
Fri Apr 20 09:46:22 EDT 2007
Thanks, Martin and Constantinos
For now I'm using mpif90, which adding the flag - mcmodel=medium did not solved the problem (the optFile I'm using is below).
Any suggestions (before moving to other compilers)?
Many thanks, Eli
FC=mpif90
F90C=mpif90
DEFINES='-DWORDLENGTH=4'
LINK='ifort'
CPP='cpp -traditional -P'
#INCLUDES='-I/usr/include/netcdf'
INCLUDES='-I/opt/netcdf-3.6.0-p1/include -I/usr/local/topspin/mpi/mpich/include'
LIBS='-L/opt/netcdf-3.6.0-p1/lib -L/usr/local/topspin/mpi/mpich/lib64 -lmpich_i -lpthread'
# Note that the -mp switch is for ieee "maintain precision" and is
# roughly equivalent to -ieee
if test "x$IEEE" = x ; then
FOPTIM='-O3 -align'
# FOPTIM=
#P3 FOPTIM=$FOPTIM' -tpp6 -xWKM'
#P4 FOPTIM=$FOPTIM' -tpp7 -xWKM'
FFLAGS='-132 -r8 -i4 -w95 -convert big_endian -assume byterecl'
else
FOPTIM='-O0 -noalign'
FFLAGS='-132 -r8 -i4 -w95 -W0 -WB -xN -pc64 -convert big_endian -assume byterecl'
#FFLAGS='-mp -132 -r8 -i4 -w95 -W0 -WB'
fi
************************************************
Eliyahu Biton
Deptartment of Environmental Sciences,
Weizmann Institute of Science,
Rehovot,76100
Israel
Tel:+972-8-9344924
Fax:+972-8-9344124
************************************************
>>> Constantinos Evangelinos <ce107 at ocean.mit.edu> 04/19/07 11:57 PM >>>
On Thu 19 Apr 2007 11:23, Martin Losch wrote:
> Hi Eli,
>
> the error message means that you are asking for too much stack
> memory. Many compilers can deal with more stack memory by specifying
> an option such as "pgf77 - mcmodel=medium".
> You'll have to find out what the option is for your compiler.
GNU, PGI, Pathscale, Absoft and Intel compilers all accept - mcmodel=medium
which has to do with the addressing mode not the stack. You have to make
sure all the libraries you link with are also compiled with - mcmodel=medium
or - fPIC.
Constantinos
--
Dr. Constantinos Evangelinos
Department of Earth, Atmospheric and Planetary Sciences
Massachusetts Institute of Technology
_______________________________________________
MITgcm- support mailing list
MITgcm- support at mitgcm.org
http://mitgcm.org/mailman/listinfo/mitgcm- support
More information about the MITgcm-support
mailing list