Open MPI logo

MTT Devel Mailing List Archives

  |   Home   |   Support   |   FAQ   |  

This web mail archive is frozen.

This page is part of a frozen web archive of this mailing list.

You can still navigate around this archive, but know that no new mails have been added to it since July of 2016.

Click here to be taken to the new web archives of this list; it includes all the mails that are in this frozen archive plus all new mails that have been sent to the list since it was migrated to the new archives.

Subject: Re: [MTT devel] Tracking process stats
From: Jeff Squyres (jsquyres_at_[hidden])
Date: 2008-04-22 17:12:42


On Apr 22, 2008, at 5:08 PM, Ethan Mallove wrote:

>> Gathering the info from any given might be a little difficult -- we'd
>> have to enlist orte somehow. The orteds could conceivably be able to
>> get this info because the MPI processes are their children, after
>> all. The orteds would then need to gather this info and make it
>> available somehow (e.g., mpirun could dump out a file, or perhaps
>> something to stdout, ...etc.).
>
> What about using Proc::ProcessTable or the "ps" command?
>
> http://www.perlmonks.org/?node_id=115098
>
> I'm guessing this would be difficult because of processes
> on remote nodes?

'zactly. I think we'd have a better chance of having ORTE report this
stuff for us, and MTT snarfing it up. Someone want to talk to Ralph?

>
> -Ethan
>
>
>>
>> On Apr 22, 2008, at 3:16 PM, Josh Hursey wrote:
>>
>>> I was thinking for very test as an extra bit or two of information.
>>> But we could try a test suite too, but we would have to think about
>>> what we would want to exercise in it.
>>>
>>> -josh
>>>
>>> On Apr 22, 2008, at 1:45 PM, Ethan Mallove wrote:
>>>
>>>> On Tue, Apr/22/2008 01:35:06PM, Josh Hursey wrote:
>>>>> On the Open MPI teleconf this morning Rich mentioned that
>>>>> he was noticing odd memory usage. It got me thinking,
>>>>> would it be useful for MTT to track important aspects of
>>>>> the process such as memory use?
>>>>>
>>>>> Just a thought. I'm not exactly sure how we would report
>>>>> or record this, but if it is useful we can file a bug for
>>>>> it :)
>>>>>
>>>>
>>>> Would this require a new "memory-usage" test suite? Or were
>>>> you thinking more along the lines of a new integer column in
>>>> the test_run table to record memory usage for *every* test
>>>> run?
>>>>
>>>> -Ethan
>>>>
>>>>
>>>>> Cheers,
>>>>> Josh
>>>>>
>>>>> _______________________________________________
>>>>> mtt-devel mailing list
>>>>> mtt-devel_at_[hidden]
>>>>> http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel
>>>> _______________________________________________
>>>> mtt-devel mailing list
>>>> mtt-devel_at_[hidden]
>>>> http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel
>>>
>>> _______________________________________________
>>> mtt-devel mailing list
>>> mtt-devel_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel
>>
>>
>> --
>> Jeff Squyres
>> Cisco Systems
>>
>> _______________________________________________
>> mtt-devel mailing list
>> mtt-devel_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel
> _______________________________________________
> mtt-devel mailing list
> mtt-devel_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel

-- 
Jeff Squyres
Cisco Systems