[mpich-discuss] installing and using MPI_PUBLISH_NAME andMPI_LOOKUP_NAME with mpich2
Dave Goodell
goodell at mcs.anl.gov
Thu Jun 16 10:16:54 CDT 2011
Frankly, I'm not sure. This is infrequently used functionality in MPICH2 (and MPI in general), so you may have encountered a bug that nobody else has hit yet. But I believe that our main hydra developer (Pavan) is looking at the issue as time permits.
You are certainly welcome to try mpd if you like, we just won't be able to help if you run into trouble. And there are many, many ways that mpd can give you problems.
-Dave
On Jun 16, 2011, at 9:21 AM CDT, Bernard Secher - SFME/LGLS wrote:
> OK Dave, so why I get these error messages with hydraserver ?
>
> bERNARD
>
> Dave Goodell a écrit :
>> Be aware that Leon's advise is exactly the opposite of what we (the MPICH2 developers) would recommend. The mpd process manager is deprecated and we really don't support it anymore.
>>
>> -Dave
>>
>> On Jun 16, 2011, at 7:40 AM CDT, Bernard Secher - SFME/LGLS wrote:
>>
>>
>>
>>> My test program work very well with opnempi, but I want to use mpich2 too, if it is possible...
>>>
>>> How can I use mpd with mpich2 ?
>>>
>>> Leon Yuhanov a écrit :
>>>
>>>
>>>> hey,ditch hydra. use mpd its easier for quick simple stuff. move onto lam-mpi or open-mpi after that.
>>>>
>>>>
>>>> Bernard Secher - SFME/LGLS
>>>>
>>>> <bernard.secher at cea.fr>
>>>>
>>>> wrote:
>>>>
>>>> When I tried to use hydra_nameserver I get the followinf error message
>>>> when I launch my application:
>>>>
>>>> [secher at is209011 essai]$ [mpiexec at is209011] HYDU_sock_read
>>>> (/data/tmplgls/secher/mpich2-1.3.2p1-src/src/pm/hydra/utils/sock/sock.c:223):
>>>> read errno (Input/output error)
>>>> [mpiexec at is209011] control_cb
>>>> (/data/tmplgls/secher/mpich2-1.3.2p1-src/src/pm/hydra/pm/pmiserv/pmiserv_cb.c:249):
>>>> assert (!closed) failed
>>>> [mpiexec at is209011] HYDT_dmxu_poll_wait_for_event
>>>> (/data/tmplgls/secher/mpich2-1.3.2p1-src/src/pm/hydra/tools/demux/demux_poll.c:77):
>>>> callback returned error status
>>>> [mpiexec at is209011] HYD_pmci_wait_for_completion
>>>> (/data/tmplgls/secher/mpich2-1.3.2p1-src/src/pm/hydra/pm/pmiserv/pmiserv_pmci.c:206):
>>>> error waiting for event
>>>> [mpiexec at is209011] main
>>>> (/data/tmplgls/secher/mpich2-1.3.2p1-src/src/pm/hydra/ui/mpich/mpiexec.c:404):
>>>> process manager error waiting for completion
>>>>
>>>> What is wrong
>>>>
>>>> My application is just a little test program. I have attached it to this
>>>> message.
>>>>
>>>> Bernard
>>>>
>>>> Pavan Balaji a écrit :
>>>>
>>>>
>>>>
>>>>
>>>>> [please keep mpich-discuss cc'ed]
>>>>>
>>>>> On 06/16/2011 02:18 AM, Bernard Secher - SFME/LGLS wrote:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> How can I use the default publish mechanism? My Publish/lookup is
>>>>>> hapenning between two different applications. Is there an option to give
>>>>>> to mpiexec for that ?
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>> Do you have a shared file-system? If yes, you don't need to modify
>>>>> anything. It should just work. If no, then read below:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> If I use the pmi option, the mpiexec -nameserver -help give the
>>>>>> following message:
>>>>>> -nameserver: Nameserver to use for publish/lookup (format is host:port)
>>>>>>
>>>>>> I don't understand what I have to do with this option.
>>>>>> First I launch the hydra_nameserver . Do I give some options to
>>>>>> hydra_nameserver ? Do I give some options then, to the two mpiexec to
>>>>>> launch my two applications?
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>> There are no options required for hydra_nameserver. The two mpiexec's
>>>>> will need the -nameserver {hostname} option, where {hostname} is the
>>>>> name of the host where the nameserver is running.
>>>>>
>>>>> -- Pavan
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>>
>>> --
>>> _\\|//_
>>> (' 0 0 ')
>>> ____ooO (_) Ooo______________________________________________________
>>> Bernard Sécher DEN/DM2S/SFME/LGLS mailto :
>>>
>>> bsecher at cea.fr
>>>
>>>
>>> CEA Saclay, Bât 454, Pièce 114 Phone : 33 (0)1 69 08 73 78
>>> 91191 Gif-sur-Yvette Cedex, France Fax : 33 (0)1 69 08 10 87
>>> ------------Oooo---------------------------------------------------
>>> oooO ( )
>>> ( ) ) /
>>> \ ( (_/
>>> \_)
>>>
>>>
>>>
>>> Ce message électronique et tous les fichiers attachés qu'il contient
>>> sont confidentiels et destinés exclusivement à l'usage de la personne
>>> à laquelle ils sont adressés. Si vous avez reçu ce message par erreur,
>>> merci d'en avertir immédiatement son émetteur et de ne pas en conserver
>>> de copie.
>>>
>>> This e-mail and any files transmitted with it are confidential and
>>> intended solely for the use of the individual to whom they are addressed.
>>> If you have received this e-mail in error please inform the sender
>>> immediately, without keeping any copy thereof.
>>>
>>>
>>> _______________________________________________
>>> mpich-discuss mailing list
>>>
>>> mpich-discuss at mcs.anl.gov
>>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>>>
>>>
>>>
>>
>> _______________________________________________
>> mpich-discuss mailing list
>>
>> mpich-discuss at mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>>
>>
>>
>
>
> --
> _\\|//_
> (' 0 0 ')
> ____ooO (_) Ooo______________________________________________________
> Bernard Sécher DEN/DM2S/SFME/LGLS mailto :
> bsecher at cea.fr
>
> CEA Saclay, Bât 454, Pièce 114 Phone : 33 (0)1 69 08 73 78
> 91191 Gif-sur-Yvette Cedex, France Fax : 33 (0)1 69 08 10 87
> ------------Oooo---------------------------------------------------
> oooO ( )
> ( ) ) /
> \ ( (_/
> \_)
>
>
>
> Ce message électronique et tous les fichiers attachés qu'il contient
> sont confidentiels et destinés exclusivement à l'usage de la personne
> à laquelle ils sont adressés. Si vous avez reçu ce message par erreur,
> merci d'en avertir immédiatement son émetteur et de ne pas en conserver
> de copie.
>
> This e-mail and any files transmitted with it are confidential and
> intended solely for the use of the individual to whom they are addressed.
> If you have received this e-mail in error please inform the sender
> immediately, without keeping any copy thereof.
>
>
> _______________________________________________
> mpich-discuss mailing list
> mpich-discuss at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
More information about the mpich-discuss
mailing list