Hi Daniela,
We don't do anything like that. If ATLAS or CMS want to run a multi-core job at RAL they just request how many cores they want, e.g. by including "( countpernode = 8 )" in their job description. This information is then passed to HTCondor (the RequestCpus attribute), resulting in multi-core jobs.
Regards,
Andrew.
________________________________
From: Testbed Support for GridPP member institutes [[log in to unmask]] on behalf of Daniela Bauer [[log in to unmask]]
Sent: Wednesday, July 12, 2017 2:59 PM
To: [log in to unmask]
Subject: multi core jobs on ARC CEs
Hi All,
currently DIRAC does not support submitting multi core jobs to ARC CEs.
When we enabled multicore jobs for CMS on our ARC CEs, we had to introcude a special file:
[root@cetest01 rte]# cat ENV/MC8
#!/bin/sh
if [ -z $1 ] || [ $1 == 0 ] ; then
# Enable Multi-core PE
export joboption_nodeproperty_0="grid8.pe<http://grid8.pe>"
fi
and CMS requires this file in their rsl.
Now poking around at other sites, it seems like this is not standard, however it looks like in our setup we definitely need it.
Does anyone know if the is just SGE and/or CMS ? (Or Imperial?)
What did other people with ARCs have to do to allow multiore jobs ?
I'm just dreading to have to setup queues by hand.
Regards,
Daniela
--
Sent from the pit of despair
-----------------------------------------------------------
[log in to unmask]<mailto:[log in to unmask]>
HEP Group/Physics Dep
Imperial College
London, SW7 2BW
Tel: +44-(0)20-75947810
http://www.hep.ph.ic.ac.uk/~dbauer/<http://www.hep.ph.ic.ac.uk/%7Edbauer/>
|