ATOMIN Cluster Info, News etc
The system is back, we have brand new debian stable system (same as shiva
cluster). With the change of system we have got a brand new cluster resource
system called slurm, read man slurm or the quick start for in:
We also have only two partitions/queues called "bigone" and "small" for the
most lazy qsub should accept old submissions scripts (with that new queue
instead of old ones), this is only a temporary measure and may not work as
expected in which case one should use native for slurm sbatch with a similar
batch files adopted from PBS to SBATCH.
Unfortunatelly in the mean time we have lost old /home filesystem, a new one
should be safe for some time.
Please send me an e-mail (Roman.Marcinek@uj.edu.pl) if you want some
software and/or libraries installed in the cluster: gnu compilers and
lapack+openblas are already there so apply for more if need be.
PLEASE be friendly to everybody and place in your SBATCH submitting script
number of machines via -N parameter and number of tasks/cpu cores via -n
one, that way jobs not filling whole machine may be running
together without degrading cluster and job performance :-)