We have SLURM 19.05. and implemented the cons_tres scheduling type. It does work only by specifying the --gpus-per-node when submitting the job. And there are many more options.
I found this presentation to be quite informative: https://slurm.schedmd.com/SLUG18/cons_tres.pdf We still have the gres configuration, users have their workload scripted and some still use sbatch with gres. Both options work. Cheers, Barbara On 8/14/19 6:06 AM, Chris Samuel wrote: > On 13/8/19 1:36 pm, Christopher Benjamin Coffey wrote: > >> Thanks for that Chris!:) > > No worries Chris! > >> Sounds like other than the new requests for gpu specifics, things >> should just work when upgrading to 19.05 as slurm is likely backwards >> compatible with the previous setup gres stuff. > > Yup, we didn't change our config for GPUs and (so far) things seem OK. > > All the best, > Chris
