[MITgcm-support] MPI on SGI (irix64_ip35)!

[Ed Hill]

On Thu, 2005-04-28 at 15:57 -0400, Van Thinh Nguyen wrote:
> 
> I tried to compile an example (eg. exp1) with -mpi option. I haven't got
> any warming. However, when i run the obtained object file (mitgcmuv) with
> command:
> 
> mpirun -np 4 ./mitgcmuv
> 
> I got an error from STDERR.* that the number of processes must equal
> nPx*nPy. I even set nPx=2, nPy=2 and changed the suitable size of number
> of tiles in SIZE.h. I also tried to use exch2 pkg by using driver.m in the
> utils/exch2/matlab_topology_generator (edited nr,ng,nb; tnx, tny), but the
> driver.m didn't work when I ran it on Matlab.
> 
> Anyone has any experience on this problem, please help me!


Hi Van Thinh,

I don't know whats going wrong for you but heres a few things to check:

 1) Can you create and successfully run simple "hello world"-type 
    MPI programs?  This is an important first thing to test since 
    you can use it to determine whether your MPI (mpich, LAM, etc.)
    install is setup and working correctly.

 2) Does the "MITgcm/verification/exp1" example run for you with 
    2 CPUs using the "MITgcm/verification/exp1/code/SIZE.h_mpi" 
    file instead of the default "SIZE.h" file?

 3) ONLY after passing both (1) + (2) above should you start making 
    changes to your setup.  Its good to make sure that the basics 
    are working before you start changing the model.

Hope that helps.

Ed

-- 
Edward H. Hill III, PhD
office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
             Cambridge, MA 02139-4307
emails:  eh3 at mit.edu                ed at eh3.com
URLs:    http://web.mit.edu/eh3/    http://eh3.com/
phone:   617-253-0098
fax:     617-253-4464