The HPC backend currently submits all jobs to one queue that is defined in the configuration of the serverSHARK. I suggest to keep this and use it as a default and add an option to the job submission interface that allows the selection of a queue by name.
Reason: mpi-short queue has a higher priority and we know that many jobs have a low execution time (limit 2h for mpi-short).
The HPC backend currently submits all jobs to one queue that is defined in the configuration of the serverSHARK. I suggest to keep this and use it as a default and add an option to the job submission interface that allows the selection of a queue by name.
Reason: mpi-short queue has a higher priority and we know that many jobs have a low execution time (limit 2h for mpi-short).