[MITgcm-support] Executable fails, message: Trace/BPT trap: 5

Martin Losch Martin.Losch at awi.de
Fri Feb 22 08:01:10 EST 2013


Nick,

what exactly is the error message? You can do: "make makefile && make CLEAN && make depend && make
and then when you start ./mitgcmuv you get Trace/BPT trap: 5?
Did you try this on a "clean" check-out on one of the verification exeriments (e.g. cvs co MITgcm_verif_basic (to have only a few verification experiments), cd MITgcm/verification; testresport -t lab_sea)

I also updated my OS to 10.8 recently, and did not have too many problems (with the MITgcm, there were tons with the OS). What is the gfortran version that you are using (I use version 4.8.0 20120930 (experimental))? Did you try without the mnc package?

Martin

On Feb 22, 2013, at 12:01 PM, "Rogan, Nicholas" <N.Rogan at liverpool.ac.uk> wrote:

> Apologies, i missed the opt file:
> 
> 
> #!/bin/bash
> 
> # $Header: /u/gcmpack/MITgcm/tools/build_options/darwin_ia32_gfortran,v 1.2 2006/12/05 22:15:42 heimbach Exp $
> #
> # tested on MacBook Pro
> # on 28-Nov-2006
> #
> # Modified for iMac9,1 OSX 10.5.8 on 11th July 2012
> #
> 
> CPP='/usr/bin/cpp -traditional -P'
> FC=gfortran
> CC=gcc
> LINK=${FC}
> NOOPTFLAGS='-O0'
> 
> ###FC_NAMEMANGLE="#define FC_NAMEMANGLE(X) _ ## X ## _"
> ###FC_NAMEMANGLE="#define FC_NAMEMANGLE(X) X"
> ###FC_NAMEMANGLE='#define FC_NAMEMANGLE(X) X ## __'
> 
> #  For IEEE, use the "-ffloat-store" option
> if test "x$IEEE" = x ; then
>     #FFLAGS='-Wunused -Wuninitialized -fsecond-underscore -fconvert=big-endian -fbacktrace'
>     FFLAGS='-Wuninitialized -fno-second-underscore -fconvert=big-endian -fbacktrace'
>     FOPTIM='-O3 -funroll-loops'
> else
>     FFLAGS='-ffloat-store -fsecond-underscore -fconvert=big-endian -fbacktrace'
>     FOPTIM='-O0'
> fi
> 
> S64='$(TOOLSDIR)/set64bitConst.sh'
> DEFINES='-DWORDLENGTH=4 -DNML_TERMINATOR'
> INCLUDEDIRS='/usr/local/netcdf/include'
> INCLUDES="-I/usr/local/netcdf/include"
> LIBS="-L/usr/local/netcdf/lib -lnetcdf -lnetcdff"
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support




More information about the MITgcm-support mailing list