[MITgcm-support] basics

[MARIO DURAN CAMEJO]

Hi Gus,

thank you very much! I will check all of that you have mentioned.

Mario

2015-06-18 11:43 GMT-07:00 Gus Correa <gus at ldeo.columbia.edu>:

> Hi Mario
>
> If you compiled the model with MPI, and your SIZE.h file is setup to
> run the model in parallel (i.e. more than one processor),
> then you cannot launch it simply with:
>
> ./mitgcmuv > output.txt
>
> Instead, you need to launch it with mpirun (or mpiexec),
> something like this:
>
> mpirun -np 16 ./mitgcmuv > output.txt
>
> where I made up 16 as the number of processors, but you need to use
> the appropriate number that matches the number of processors
> that you set up in SIZE.h.
>
> **
>
> In addition, you may also have an environment problem,
> maybe a mixup of different MPI implementations.
>
> Make sure the MPI used for compilation and linking
> (check your build option file in tools/build_options/)
> is the same used to launch the mitgcmuv executable
> (through mpiexec or mpirun).
>
> It may help also if you check the output of:
>
> 'which mpif77' (or 'which mpif90', or 'which mpifortran')
>
> and
>
> 'which mpirun' (or 'which mpiexec')
>
> Very often people have a variety of MPI flavors installed (MPICH, OpenMPI,
> even duplicate ancient versions) and their PATH and LD_LIBRARY_PATH
> environment variables are set up inconsistently,
> which can lead to the mixup.
> The runtime environment must match the compilation/linking environment.
>
> I hope this helps,
> Gus Correa
>
> On 06/18/2015 01:27 PM, Edwards, Erick (LCDR) wrote:
>
>> Mario,
>>
>> I have some batch files and other information to help out.  Let’s meet
>> this morning in my office.
>>
>> Erick
>>
>> *From:*MARIO DURAN CAMEJO [mailto:mdcamejo at ing.uc3m.es]
>> *Sent:* Thursday, June 18, 2015 7:19 AM
>> *To:* MITgcm-support at mitgcm.org
>> *Subject:* [MITgcm-support] basics
>>
>> Hello,
>>
>> I'm just beginning with MITgcm and I have a very simple question.
>>
>> I have built my executable in the build directory and made a copy to
>> input, but when I enter ./mitgcmuv > output.txt I get an error like this:
>>
>> [unset]: aborting job:
>>
>> Fatal error in MPI_Init:
>>
>> Other MPI error, error stack:
>>
>> MPIR_Init_thread (436)..........:
>>
>> MPID_Init (371).....................: channel initialization failed
>>
>> MPIDI_CH3_Init (335)............:
>>
>> MPIDI_CH3I_SMP_Init (1723):  LiMIC2 device does not exist: No such file
>> or directory
>>
>> Does someone know what is going on?
>>
>> Thank you very much!
>>
>> Mario
>>
>> --
>>
>> MARIO DURAN CAMEJO
>> Universidad Carlos III de Madrid
>>
>>
>>
>> _______________________________________________
>> 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
>



-- 
MARIO DURAN CAMEJO
Universidad Carlos III de Madrid
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20150618/cd6b240e/attachment.htm>