Hi,

Am 30.03.2014 um 15:03 schrieb Edrisse Chermak:

> Dear Grid Engine developers and users,
> 
> I would like to bind processes to a sequence of cores or to a single
> socket.
> I tried to constrain my jobs to run on 8 cores within 1 socket :
> 
> qsub -S /bin/bash -N test -cwd -V -pe smp 8 -binding set striding:1:8 -q
> 1day.q job.sh
> 
> Unfortunately, I see processes jumping from cores to other cores during
> the job.
> I perhaps missed something, could you please advise me ?
> 
> Best Regards,
> Edrisse
> 
> Notes :  - I use Grid Engine 2011.1
>         - I set parallel environment on pe_slots
>         - My nodes architecture is 64 cores on 4 sockets :

did you enable binding in execd_params in SGE's configuration?

$ qconf -sconf
...
execd_params                 ENABLE_ADDGRP_KILL=TRUE ENABLE_BINDING=TRUE

Or do you mean that they are jumping inside the set of granted cores?

The allocation is also shown in /proc/<PID>/status

-- Reuti

> $loadcheck
> arch            linux-x64
> num_proc        64
> m_socket        4
> m_core          64
> m_topology
> SCCCCCCCCCCCCCCCCSCCCCCCCCCCCCCCCCSCCCCCCCCCCCCCCCCSCCCCCCCCCCCCCCCC
> 
> ________________________________
> 
> This message and its contents including attachments are intended solely for 
> the original recipient. If you are not the intended recipient or have 
> received this message in error, please notify me immediately and delete this 
> message from your computer system. Any unauthorized use or distribution is 
> prohibited. Please consider the environment before printing this email.
> 
> _______________________________________________
> users mailing list
> [email protected]
> https://gridengine.org/mailman/listinfo/users


_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users

Reply via email to