Horus night-time queue description
Horus is managed by SLURM. Detailed information about SLURM is available.
A few commands will get you a long way down the road. However, to know more, just look at the man page
squeue: | view information about jobs located in the SLURM scheduling queu |
sinfo: | view information about SLURM nodes and partitions |
srun: | run serial/parallel jobs |
The night-time queue called "cpunight" is open from 20:00 until 8:00am every day. Jobs running at 8:00am will be allowed to continue until completion.
You may submit any single-node job to the queue with the command:
srun --batch --partition=cpunight --share --verbose --mail-type=ALL --mail-user=emailaddress@yourprovider.ch /users/yourid/job.sh
where job.sh contains a simple command to execute your job, as if running it from the console. Standard output will be redirected (by default) to a file called slurm-jobid.out.
PBS Pro System Description
The HORUS frontend may also access the whole PRE/POST Processing RIGI computing cluster resources via the PBS Professional V 9.1.0 Client application.
A standard Module is preloaded into the user's default environment on the frontend horus.cscs.ch (pbspro/pbspro-9.1.0).
For further details about job submission, see the official documentation available for the RIGI cluster.