srun：error：

[wangzc]

hi，
No prompt error！
All log information followes:
******************************************************************************
Model Input Parameters: ROMS/TOMS version 3.6
Tuesday - September 17, 2013 - 7:54:04 PM
-----------------------------------------------------------------------------
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
oceanM 00000000006EDB5F Unknown Unknown Unknown
oceanM 00000000005077D6 Unknown Unknown Unknown
oceanM 0000000000476618 Unknown Unknown Unknown
oceanM 000000000040B0CE Unknown Unknown Unknown
oceanM 0000000000405CF1 Unknown Unknown Unknown
oceanM 0000000000405BE2 Unknown Unknown Unknown
libc.so.6 0000002A95C823FB Unknown Unknown Unknown
oceanM 0000000000405B2A Unknown Unknown Unknown
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
oceanM 00000000006EDB5F Unknown Unknown Unknown
oceanM 00000000005077D6 Unknown Unknown Unknown
oceanM 0000000000476618 Unknown Unknown Unknown
oceanM 000000000040B0CE Unknown Unknown Unknown
oceanM 0000000000405CF1 Unknown Unknown Unknown
oceanM 0000000000405BE2 Unknown Unknown Unknown
libc.so.6 0000002A95C823FB Unknown Unknown Unknown
oceanM 0000000000405B2A Unknown Unknown Unknown
srun: error: xcn053: task1: Exited with exit code 174
srun: error: xcn053: task0: Exited with exit code 174
********************************************************************************
DEBUGGING is actived in build.bash, but no more information. why?
--zongchen

[kate]

Did you active DEBUGGING or USE_DEBUG? Did you recompile?

[wangzc]

hi Kate
DEBUGGING and DIS4DVAR is actived in build.bash, and recompiling is ok. H-file follows:
********************************************************************************
*/
/*
**-----------------------------------------------------------------------------
** Nonlinear basic state settings.
**-----------------------------------------------------------------------------
*/


#define UV_ADV
#define DJ_GRADPS
#define UV_COR
#define UV_QDRAG
#define UV_VIS2
#define CURVGRID
#define PROFILE
#define SPLINES
#define MASKING

/*
** Common options to all 4DVAR algorithms.
*/

#if defined ARRAY_MODES || defined CLIPPING || \
defined IS4DVAR || defined IS4DVAR_SENSITIVITY || \
defined W4DPSAS || defined W4DPSAS_SENSITIVITY || \
defined W4DVAR || defined W4DVAR_SENSITIVITY
# define ADJUST_WSTRESS
# define FORWARD_WRITE
# define FORWARD_READ
# define OUT_DOUBLE
#endif

******************************************************************

Because of computing cost and memery limits, i shut down all cpp options connected with SOLVE3D.
--zongchen

[arango]

That's your problem. The 4D-Var algorithms are primarily for 3D applications using SOLVE3D. I won't guarantee the shallow-water (only 2D) equations are running with 4D-Var nowadays. We have that working in the early years for debugging purposes only.

When I get this kind of error, I always compile and run the model in serial and not in parallel. This will help with the debugging. If you have limited memory in your computer, you should not run any of the adjoint-based algorithms. They are memory intensive and require from 4-5 times the memory of just running the nonlinear model. You can only run very small 4D-Var problems in a small computers (like a laptop, WC13).

You cannot run yet ROMS in a smartphone ...

[wangzc]

hi arango，
In which my 2D_ROMS run is a cluster, and i ask for a 100G space. I think one week simulation is able to satified.
U say 'I won't guarantee the shallow-water (only 2D) equations are running with 4D-Var nowadays.'. Do u mean if i want to do the way i want, i have to modify the code and debug time after time? If so, maybe i give up the experiment temporarily until my 3D_ROMS is well configured.
--zongchen