On 16/07/16 05:13, Steven Lo wrote: > One of the feature that we have implemented is queue routing where > all the jobs are submitted to productionQ queue and it will route to > the appropriate queue automatically.
One major difference with Slurm is that you can submit to *multiple* partitions at the same time and any that you do not have access to will be automatically pruned (as long as at least one is left, otherwise the job is rejected). Your job will then run in the first eligible partition that it can. There is a sample "all_partitions" job submit plugin provided. http://slurm.schedmd.com/job_submit_plugins.html So whilst Torque is a classically mechanical system in the physics sense Slurm is more quantum - your job exists in many states until it can start running and then it collapses into its final state. :-) This is something that I'm still wrapping my head around having also come from a Torque/OpenPBS background and so we don't take advantage of it ourselves yet, but I have plans... Best of luck! Chris -- Christopher Samuel Senior Systems Administrator VLSCI - Victorian Life Sciences Computation Initiative Email: sam...@unimelb.edu.au Phone: +61 (0)3 903 55545 http://www.vlsci.org.au/ http://twitter.com/vlsci
