[MITgcm-support] forrtl: severe (36) with tile file

Jean-Michel Campin jmc at ocean.mit.edu
Mon Aug 26 18:48:24 EDT 2013


Hi Neil,

This is the same error you reported earlier (Aug 20) right ?

This error:
> forrtl: severe (36): attempt to access non-existent record, unit 9, file /usr/data/nppatel/MITgcm_c64g/verification/cube_test/input/tile001.mitgrid
is quite common (might find similar report in the support archive) 
and indicates that the executable (in ini_curvilinear_grid.F)
is trying to read a value that is beyond the end of the file.
I would recommend to check SIZE.h (and, if provided, data.exch2)
 and the size of these tile00[1-6].mitgrid files.
 The size of each file (given by "ls -l") should be:
   = 8*(Nc+1)*(Nc+1)*16 
 where Nc is the size of 1 (squared) face (e.g., Nc=32 for cs-32 grid)
  and 16 correspond to the number of different fields.
Also remember that they are double prec (the "8" factor), even if 
  readBinaryPrec=32.

Cheers,
Jean-Michel

On Mon, Aug 26, 2013 at 06:03:08PM -0400, Neil Patel wrote:
> I'm using a cube face tile grid, with six files that describe the grid system which are named like this:
> 
> tile001.mitgrid 
> 
> When I run mitgcm, it crashes when it gets to the point where it reads tile001.mitgrid with this error message:
> 
> forrtl: severe (36): attempt to access non-existent record, unit 9, file /usr/data/nppatel/MITgcm_c64g/verification/cube_test/input/tile001.mitgrid
> 
> The file is there, it crashes while reading it I think. Anyone have a clue on the source of the error? Thanks! Here's my build file:
> 
> FC=/opt/intel/bin/ifort
> F90C=/opt/intel/bin/ifort
> qCC=cc
> LINK='/opt/intel/bin/ifort'
> 
> DEFINES='-DALLOW_USE_MPI -DALWAYS_USE_MPI -DWORDLENGTH=4'
> CPP='cpp -P -traditional -D_d=D'
> F90FIXEDFORMAT='-fixed -Tf'
> EXTENDED_SRC_FLAG='-132'
> 
> MCMODEL='-mcmodel=medium -shared-intel'
> FFLAGS="$FFLAGS -W0 -WB -convert big_endian -assume byterecl $MCMODEL"
> 
> —Neil
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support



More information about the MITgcm-support mailing list