[MITgcm-support] How to create .nc file- (How to search the MITgcm list)

[Abbas Dorostkar]

Hi Martin and Randhir
I usually use Google search tool to browse the support list. Click on
the advanced search option on the google page, put this link"
http://forge.csail.mit.edu/pipermail/mitgcm-support/" in the "Search
within a site or domain box". In this way, we can search the list
easily.
Abbas








-----Original Message-----
From: mitgcm-support-bounces at mitgcm.org
[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Martin Losch
Sent: May 28, 2008 3:12 AM
To: mitgcm-support at mitgcm.org
Subject: Re: [MITgcm-support] How to create .nc file

Randhir,

that's a system error message and has to do with you MPI environment  
and how it is invoked. Talk to you system administrator or someone  
who know how to run mpi programs on your linux cluster.

On the other hand, for exp2, you don't really need more than one cpu  
and using MPI may be an overkill. To compile the model without MPI  
use a different opt file, eg. linux_ia32_g77, whatever better  
compiler you have available; g77 is not the most efficient compiler  
(although the model runs fine with it).

To get netcdf output you need
1. to make sure that the netcdf libraries are installed on your system
2. to make sure your opt files points to the include and library  
files (INCLUDES and LIBS variables)
3. to activate the package mnc in a file code/packages.conf and use  
the appropriate flags in data.pkg and data.mnc.
Please see an example in the verification experiment that does so,  
e.g. "tutorial_global_oce_latlon" is close to exp2 in terms of domain  
and has an appropriate packages.conf, the only thing you'd need to do  
is turn on the mnc package in data.pkg (uncomment #useMNC=.true.)

This question has been answered a million times on the list (however,  
since the search tool does not really work, you wouldn't be able to  
find out about that) and it's also described in the manual, here:  
<http://mitgcm.org/r2_manual/latest/online_documents/node266.html>

Martin

On 28 May 2008, at 05:32, randhir singh wrote:

> Sir,
> I have sucessfully created the mitgcmuv file for exp2. Now please  
> guide me through the steps by which I can create the .nc file to  
> proceed further in the analysis of the experiment using grads in  
> windows xp.
> When I execute the mpirun -np 1 ./mitgcmuv command I get the  
> following error message, the statement is as follows.
>
> " It seems that [atleast] one of the processes that was started  
> with mpirun did not invoke MPI_INIT before quitting........mpirun  
> can *only*  be used with mpi programs (i.e programs that invoke  
> MPI_INIT and MPI_FINALIZE) " Now guide me how to remove this error.
> I am using the optfile as /linux_ia32_g77+mpi_cg01.
>  I have copied the MITgcm  experiment on the Red Hat Linux server  
> and use telnet on windows xp to get connected to the linux server.
>
> Regards,
> Randhir Singh.
>
>
>
>
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support

_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org
http://mitgcm.org/mailman/listinfo/mitgcm-support