Job Types

The clusters are designed for different work loads, and each cluster has several types of jobs. This page gives an overview of the available job types on each cluster, and their main characteristics. See the sub pages of each cluster for a more detailed description of the job types and their purposes.

You should always choose the job type that fulfils the resource requirements of your job best without being excessive as this ensures the highest possible priority and therefore the shortest queuing time. If for example your job on Saga needs 20GB memory and 5 CPUs for 5 days, you should choose a normal instead of a bigmem job. If it instead needs 200GB memory you should use bigmem.

Betzy

Name	Description	Job limits	Max walltime	Priority
normal	default job type	4–512 nodes	4 days	normal
accel	jobs needing GPUs		7 days	normal
preproc	pre-/postprocessing jobs	1–16 units	1 day	normal
devel	development jobs (compiling, testing)[1]	1–4 nodes	60 mins	high

For jobs that don’t request GPUs or much memory, the “units” of accel or preproc jobs are simply the number of cpus the job requests. For other jobs, see Projects and accounting for how the units are calculated.

Betzy Job Types.

Fram

Name	Description	Job limits	Max walltime	Priority
normal	default job type	1–32 nodes	7 days	normal
bigmem	jobs needing more memory		14 days	normal
devel	development jobs (compiling, testing)	1–8 nodes	30 mins	high
short	short jobs	1–10 nodes	2 hours	high
optimist	jobs w/checkpointing, or very short jobs	1–32 nodes	see details	low

Fram Job Types.

Saga

Name	Description	Job limits	Max walltime	Priority
normal	default job type	1–256 units	7 days	normal
bigmem	jobs needing more memory	1–256 units	14 days	normal
hugemem	jobs needing even more memory	1–256 units	14 days	normal
accel	jobs needing P100 GPUs	1–256 units	14 days	normal
a100	jobs needing A100 GPUs	1–256 units	14 days	normal
devel	development jobs (compiling, testing)[2]	1–128 units	2 hours	high
optimist	jobs w/checkpointing, or very short jobs	1–256 units	see details	low

For jobs that don’t request GPUs or much memory, the “units” on Saga are simply the number of cpus the job requests. For other jobs, see Projects and accounting for how the units are calculated.

Saga Job Types.