Donor grid identification-assignment problem

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
petalas
Posts: 12
Joined: Mon Jul 04, 2016 1:31 pm
Location: University of the Aegean, GR

Donor grid identification-assignment problem

#1 Unread post by petalas »

Hi, I was wondering if anybody came across this before. In my nested application, it seems like the code is assigning the coarse grid as 00 instead of 01. Also the timestep ratio is zero instead of 3.

Code: Select all

 Refined Nested Grid(s) Information: 
 ==================================
 Refined    Donor   Refined   Timestep   Refined
  Grid      Grid    Scale      Ratio   Timesteps
    02        00       03      0.00000      03
 WARNING: Usually the number of Refined Timesteps must be the same
          as the Refined Scale for numerical stability.
The run exits after the first couple of steps with


#3 0x48BE02 in __nesting_mod_MOD_nesting
#4 0x45CAE8 in main3d_
#5 0x4035D4 in MAIN__ at master.f90:?
--------------------------------------------------------------------------
mpiexec noticed that process rank 2 with PID 4067 on node stamlab-500-165ev exited on signal 11 (Segmentation fault).
--------------------------------------------------------------------------


and the output of the debugger is

Code: Select all

 -----------------------------------------------------------------------------
  At line 706 of file metrics.f90
Fortran runtime error: Index '0' of dimension 1 of array 'dt' below lower bound of 1
I'm using coarse2fine-contact to make the fine grid.
Any hints?
Last edited by petalas on Thu May 04, 2017 8:51 am, edited 1 time in total.

User avatar
wilkin
Posts: 875
Joined: Mon Apr 28, 2003 5:44 pm
Location: Rutgers University
Contact:

Re: Donor grid identification-assignment problem

#2 Unread post by wilkin »

INP_PAR - Unable to open ROMS/TOMS input script file.
In distributed-memory applications, the input
script file is processed in parallel. The Unix
routine GETARG is used to get script file name.
For example, in MPI applications make sure that
command line is something like:
mpirun -np 4 ocean ocean.in
and not
mpirun -np 4 ocean < ocean.in
The error message indicates ROMS never read ocean.in so does not know your time steps.

Did you do this:

Code: Select all

mpirun -np 4 ocean ocean.in
or this:

Code: Select all

mpirun -np 4 ocean < ocean.in
John Wilkin: DMCS Rutgers University
71 Dudley Rd, New Brunswick, NJ 08901-8521, USA. ph: 609-630-0559 jwilkin@rutgers.edu

petalas
Posts: 12
Joined: Mon Jul 04, 2016 1:31 pm
Location: University of the Aegean, GR

Re: Donor grid identification-assignment problem

#3 Unread post by petalas »

Thanks for the answer. I'm sorry... I miss-copy-pasted the wrong debug. I edited my post with the correct debug output.

petalas
Posts: 12
Joined: Mon Jul 04, 2016 1:31 pm
Location: University of the Aegean, GR

Re: Donor grid identification-assignment problem

#4 Unread post by petalas »

Moreover, when I run the exact same model set up with a refinement factor of 7 for the fine grid, the model runs.
The problem is that I get this error when I try it with ref. factor 3 or 5.

petalas
Posts: 12
Joined: Mon Jul 04, 2016 1:31 pm
Location: University of the Aegean, GR

Re: Donor grid identification-assignment problem

#5 Unread post by petalas »

I managed to solve this by changing slightly the borders of the refined grid, in coarse2fine.
It seems that the complex geometry of the Donor grid interferes with the refined, as I illustrate in the graph.

Stamatis
Attachments
nesting_configuration.png

Post Reply