SLURM Partition Limits¶
Jobs can run on specific HiPerGator servers or nodes based on hardware requirements. Different sets of hardware configurations are presented as SLURM partitions. Please see our Available Node Features page for more details.
Time Limits¶
Partitions have individual time limits. These limits can be set in your job script using the following configuration:
SBATCH --time=4-00:00:00 # Walltime in hh:mm:ss or d-hh:mm:ss
Interactive and GUI Sessions¶
| Partition | Default Time Limit | Max Time Limit |
|---|---|---|
| hpg-dev | 10 Minutes | 12 Hours |
| hwgui | 10 Minutes | 4 Days |
Compute Partitions¶
| Partition | Default Time Limit | Max Time Limit |
|---|---|---|
| hpg-default | 10 Minutes | 31 Days |
| bigmem | 10 Minutes | 31 Days |
| hpg-milan | 10 Minutes | 31 Days |
| hpg-turin | 10 Minutes | 14 Days |
| hpg-b200 | 10 Minutes | 14 Days |
QOS¶
| Partition | Default Time Limit | Max Time Limit |
|---|---|---|
| Burst QOS | 10 Minutes | 4 Days |
| Investment QOS | 10 Minutes | 31 Days |
Note
There is no burst QOS for the gpu partitions.
OpenOnDemand¶
- GPU jobs are limited to 12 hours.
- All other jobs follow their corresponding partition limits.
Jupyter¶
- JupyterHub sessions are preset with individual limits in the 'Server Options' menu shown below:
