Skip to content


Default queues

Queues are a basic concept of PBS. When submitting a job, the user normally does not have to specify any queue. For every PBS server, there exists a default queue, namely


where the job goes if not specified otherwise.

Specific queues

In some cases, it is possible or even recommended that the user choses a particular queue. This is the case when the job is characterized not by walltime as a major limiting factor, but by another resource - memory, GPU usage etc.

Queue name Description Jobs requiring at least 1 GPU, up to 24 hours walltime Jobs requiring at least 1 GPU, up to 2 weeks walltime Jobs requiring 500 GB or more, up to 1 week walltime Jobs requiring at least 1 GPU, up to 24 hours walltime Jobs optimized to run on Intel Phi architecture Jobs requiring >100 CPUs OR >%)) GB of memory

Queue info by qstat

The qstat command provides info about queues and jobs.


qstat -q  # get list of queues and their properties on current PBS server
qstat -Q  # dtto, different format
qstat -q # list queues on PBS server
qstat -q # list queues on all servers

To see details for a selected queue, use:

qstat -Q -f queue_name@server_name

Example of output of qstat -Q -f

(BUSTER)user123@skirit:~$ qstat -Q -f
Queue: gpu_long
    queue_type = Execution
    Priority = 66
    total_jobs = 24
    state_count = Transit:0 Queued:14 Held:0 Waiting:0 Running:10 Exiting:0 Beg
    max_queued = [u:PBS_GENERIC=2000]
    resources_max.ngpus = 99
    resources_max.walltime = 336:00:00
    resources_min.mem = 50mb
    resources_min.ncpus = 1
    resources_min.ngpus = 1
    resources_default.ngpus = 0
    comment = Queue for long time computations on GPU|Fronta pro dlouhodob\u00e
    9 v\u00fdpo\u010dty na GPU
    default_chunk.queue_list = q_gpu_long
    resources_assigned.mem = 1760gb
    resources_assigned.mpiprocs = 124
    resources_assigned.ncpus = 124
    resources_assigned.nodect = 10
    max_run_res.ncpus = [u:PBS_GENERIC=200]
    backfill_depth = 2
    enabled = True
    started = True

In this particular queue, no jobs can run unless they require at least 1 GPU (resources_min.ngpus).