Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] Changing the MPIRUN/MPIEXEC semantics
From: Ralph Castain (rhc_at_[hidden])
Date: 2009-07-03 12:58:18


app0 will have ranks 0-2, app2 will have rank 3-(2+x), app3 will have
rank....you get the picture.

On Jul 3, 2009, at 10:14 AM, Simone Pellegrini wrote:

> Ralph Castain wrote:
>> Sure:
>>
>> mpirun --np 3 mpi_app1 "app1_args" : -np x mpi_app2 "app2_args" : -
>> np y mpi_app3 "app3_args"
>>
> Nice, but what the implication with the process rank?
>
> Can I assume that app1 will have rank 0, app2 rank 1 and app3 rank
> 3? or there are no assumption that can me made?
>
> thanks again, regards Simone
>>
>> On Jul 3, 2009, at 9:36 AM, Simone Pellegrini wrote:
>>
>>> Dear all,
>>> current implementation of mpirun starts the executable in
>>> different nodes. For some reason I need to start different MPI
>>> applications across nodes and I want to use MPI to communicate
>>> among these applications. For short I want to breakdown the SPMD
>>> model, something like:
>>>
>>> mpirun --np 3 mpi_app1 "app1_args", mpi_app2, "app2_args",
>>> mpi_app3, "app3_args"
>>>
>>> Is it possible to have this kind of behavior with the current
>>> implementation of mpirun?
>>> if not... it's implement it via changing the mpirun and/or mpiexec
>>> routines or does it implies radical changes to the entire Open MPI
>>> library?
>>>
>>> regards, Simone
>>> _______________________________________________
>>> users mailing list
>>> users_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>
>> _______________________________________________
>> users mailing list
>> users_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>
> _______________________________________________
> users mailing list
> users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/users