[mpich-discuss] Issue with MPI_Comm_spawn_multiple

Seo, Sangmin sseo at anl.gov
Fri Oct 9 22:47:07 CDT 2015


Hi Toni,

We figured out the reason (the spawned processes hang because when the "hosts" info keys for two commands have the same value (e.g., localhost) two processes get the same rank.), but it seems it will take some time to fix this. Pavan will rewrite the problematic part in hydra.

I’ve created a ticket (https://trac.mpich.org/projects/mpich/ticket/2308) for this issue and cc’ed you in the ticket.

Regards,
Sangmin


On Oct 8, 2015, at 10:17 AM, Antonio J. Peña <antonio.pena at bsc.es<mailto:antonio.pena at bsc.es>> wrote:


Thanks a lot Sangmin!


On 10/08/2015 03:43 PM, Seo, Sangmin wrote:
Hi Toni,

I confirmed the problem and am looking into it. I will get back to you after I resolve it.

Regards,
Sangmin


On Oct 6, 2015, at 7:23 AM, Antonio J. Peña <antonio.pena at bsc.es<mailto:antonio.pena at bsc.es>> wrote:


Right, but I do need to use the hosts key to get control of where the processes are spawned. Any clue?


On 10/02/2015 07:32 PM, Thakur, Rajeev wrote:
Looks like it is related to the hosts key. Works without that.

Rajeev

On Oct 2, 2015, at 10:34 AM, Antonio J. Peña <antonio.pena at bsc.es<mailto:antonio.pena at bsc.es>> wrote:

Hi folks!

I've been facing an issue with MPI_Comm_spawn_multiple when all the
following conditions are met:

 1. 'count' is > 1
 2. Any value of 'array_of_maxprocs' is > 1
 3. Setting the "hosts" info key

What I'm finding is that the spawned processes hang at MPI_Init. Find
attached a test case. You can reproduce my issue with -np 1.

Is there anything I'm doing wrong? I'm using MPICH master at 1e6c4d8. I
haven't found a pending ticket, xfail test, or test case covering this
situation.

All the best,
  Toni

[P.S.: It still feels weird to be on the other side of discuss at mpich.org<mailto:discuss at mpich.org>]

--
Antonio J. Peña
Senior Researcher
Barcelona Supercomputing Center



WARNING / LEGAL TEXT: This message is intended only for the use of the
individual or entity to which it is addressed and may contain
information which is privileged, confidential, proprietary, or exempt
from disclosure under applicable law. If you are not the intended
recipient or the person responsible for delivering the message to the
intended recipient, you are strictly prohibited from disclosing,
distributing, copying, or in any way using this message. If you have
received this communication in error, please notify the sender and
destroy and delete any copies you may have received.

http://www.bsc.es/disclaimer<spawn_multiple.c>_______________________________________________
discuss mailing list     discuss at mpich.org<mailto:discuss at mpich.org>
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss
_______________________________________________
discuss mailing list     discuss at mpich.org<mailto:discuss at mpich.org>
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss
--
Antonio J. Peña
Senior Researcher
Barcelona Supercomputing Center


WARNING / LEGAL TEXT: This message is intended only for the use of the
individual or entity to which it is addressed and may contain
information which is privileged, confidential, proprietary, or exempt
from disclosure under applicable law. If you are not the intended
recipient or the person responsible for delivering the message to the
intended recipient, you are strictly prohibited from disclosing,
distributing, copying, or in any way using this message. If you have
received this communication in error, please notify the sender and
destroy and delete any copies you may have received.

http://www.bsc.es/disclaimer
_______________________________________________
discuss mailing list     discuss at mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss
_______________________________________________
discuss mailing list     discuss at mpich.org<mailto:discuss at mpich.org>
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss

--
Antonio J. Peña
Senior Researcher
Barcelona Supercomputing Center


WARNING / LEGAL TEXT: This message is intended only for the use of the
individual or entity to which it is addressed and may contain
information which is privileged, confidential, proprietary, or exempt
from disclosure under applicable law. If you are not the intended
recipient or the person responsible for delivering the message to the
intended recipient, you are strictly prohibited from disclosing,
distributing, copying, or in any way using this message. If you have
received this communication in error, please notify the sender and
destroy and delete any copies you may have received.

http://www.bsc.es/disclaimer
_______________________________________________
discuss mailing list     discuss at mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20151010/99653f67/attachment.html>
-------------- next part --------------
_______________________________________________
discuss mailing list     discuss at mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss


More information about the discuss mailing list