One thing we do is in 'qconf -ssconf' load_formula slots
That will assign small jobs (e.g. one slot jobs) to hosts with the highest number of slots already used. So on an empty cluster, you'd see the small jobs fill up a node at at time, while leaving the empty nodes empty. Which is the opposive of the default behavior (load_formula load) which schedules new small jobs onto emptiest nodes.
Of course, if there are lots of small jobs still waiting in the queue, the cluster will fill up anyway.
But it's an easy change to try, IIRC no need to restart anything. http://arc.liv.ac.uk/SGE/howto/sge-configs.html#_fill_up_hosts_a_id_fillup_a On 05/01/2015 07:39 PM, Mike Hanby wrote:
Howdy, I'm wondering if anyone in the SGE community has any tips on how to accomplish this on an SGE 6.2u5p2? We would like to improve the balance competing of application profiles currently active on the cluster, in particular, we need to balance queue wait times between jobs that require many cores via MPI and those that use only a single core. Currently, MPI and SMP jobs tend to wait longer (the wait time goes up as the number of slots requested) than single slot serial jobs. We have a fair share policy in place, but even with that, serial users starve out the large # of slot MPI users, especially large memory MPI jobs (ex: 64 slots...
-- Alex Chekholko [email protected] _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
