[ExM Users] turbine-aprun-run.zsh node/proc confusion
Tim Armstrong
tim.g.armstrong at gmail.com
Fri Jun 27 12:32:27 CDT 2014
-n is the number of processes total, not number of nodes. If you want 6
processes on one node, set PPN=6 and -n 6, or if you want 6 nodes with 24
processes each, set PPN=24 and -n $((24 * 6))
We may need more documentation here.
- Tim
On Fri, Jun 27, 2014 at 11:50 AM, Ketan Maheshwari <ketan at mcs.anl.gov>
wrote:
> Setting PPN=24 on Hopper and running:
>
> turbine-aprun-run.zsh -n 6 aprog.tcl
>
> gives:
>
> aprun: -N cannot exceed -n
>
>
>
>
> On Fri, Jun 27, 2014 at 11:18 AM, Tim Armstrong <tim.g.armstrong at gmail.com
> > wrote:
>
>> The PPN environment variable will control that.
>>
>>
>> On Fri, Jun 27, 2014 at 11:14 AM, Ketan Maheshwari <ketan at mcs.anl.gov>
>> wrote:
>>
>>> Hi,
>>>
>>> Currently, when specifying n to say 4 to turbine-aprun-run.zsh, it
>>> creates a jobdef with 4 nodes but runs only 4 procs on these nodes wasting
>>> c-1 cores per node. Is there a way to specify the corespernode for target
>>> architecture and let the aprun script pick that up and launch appropriate
>>> number of procs?
>>>
>>> Thanks,
>>> Ketan
>>>
>>> _______________________________________________
>>> ExM-user mailing list
>>> ExM-user at lists.mcs.anl.gov
>>> https://lists.mcs.anl.gov/mailman/listinfo/exm-user
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/exm-user/attachments/20140627/1a5793ae/attachment.html>
More information about the ExM-user
mailing list