[MITgcm-support] Limitation for Tile's dimensions

Eli Biton eliyahub at wisemail.weizmann.ac.il
Fri Apr 20 12:03:31 EDT 2007


Hi Constantinos

The compilation error message (when using 36 CPU's) is:

sigreg.o(.text+0xe): In function `killhandler':
: relocation truncated to fit: R_X86_64_PC32 ip
sigreg.o(.text+0x15): In function `killhandler':
: relocation truncated to fit: R_X86_64_PC32 ip
sigreg.o(.text+0x37): In function `sigreg_':
: relocation truncated to fit: R_X86_64_PC32 ip
/opt/netcdf-3.6.0-p1/lib/libnetcdf.a(v2i.o)(.text+0x25f): In function `NCadvise':
/opt/netcdf-3.6.0-p1/src/libsrc/v2i.c:145: relocation truncated to fit: R_X86_64_PC32 cdf_routine_name
gmake: *** [mitgcmuv] Error 1

The optFile used is:

FC=mpif90
F90C=mpif90
DEFINES='-DWORDLENGTH=4'
LINK='ifort -i-dynamic'
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 -mcmodel=large'
else
    FOPTIM='-O0 -noalign'
    FFLAGS='-132 -r8 -i4 -w95 -W0 -WB -xN -pc64 -convert big_endian -assume byterecl -mcmodel=large'
   #FFLAGS='-mp -132 -r8 -i4 -w95 -W0 -WB'
fi


And value's parameters in SIZE.h are: 

     &           sNx = 162,
     &           sNy = 452,
     &           OLx =   3,
     &           OLy =   3,
     &           nSx =   1,
     &           nSy =   1,
     &           nPx =   4,
     &           nPy =   9,
     &           Nx  = sNx*nSx*nPx,
     &           Ny  = sNy*nSy*nPy,
     &           Nr  =  38)

Setting 72 CPU's in SIZE.h as follow leads to normal ended compilation:

     &           sNx = 324,
     &           sNy = 113,
     &           OLx =    3,
     &           OLy =    3,
     &           nSx =     1,
     &           nSy =     1,
     &           nPx =     2,
     &           nPy =   36,
     &           Nx  = sNx*nSx*nPx,
     &           Ny  = sNy*nSy*nPy,
     &           Nr  =  38)

thanks again, Eli
(I will not be near my desktop until tomorrow) 

 


************************************************
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> 04/20/07 6:26 PM >>> 
On Fri 20 Apr 2007 11:13, Eli Biton wrote:
> compilation with  -  mcmodel=medium (or -  mcmodel=large), while setting 32
> CPU's was lead to the same error message, but was good for 72 CPU's. I will
> probably will need to reduced resolution. Many Thanks, Eli

This does not make much sense. Can you please send the complete error message 
list at link- time? 

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