PC² JupyterHub
The PC² JupyterHub can be reached here and is provided by Paderborn Center for Parallel Computing, University of Paderborn.
Short Facts
- Service URL
- Target group
-
All users of a computing time project at PC²
- Login process
-
Login via LDAP
- Support
- Documentation
-
https://pc2.uni-paderborn.de/go/jh
Features
PC² JupyterHub offers:
-
Version(s): JupyterHub 3.1.0, JupyterLab 3.6
-
Temporarly install new packages by the user
-
Shared folder exists
-
Persistent storage
-
Pre-Set Environments
-
Course functionality
-
User home directory for e.g. user-specific packages
Programming Languages:
-
Python
-
Julia
-
R
-
jupyterlab-nvdashboard
-
jupyterlab-lmod
-
jupyterlab-desktop
-
jupyterlab-plotly
-
jupyterlab-pygments
-
Remote Desktop (noVNC)
-
Full offered software on clusters with graphical interface (https://pc2.uni-paderborn.de/go/software)
Resources
PC² JupyterHub provides Jupyter servers within the following resource limits:
| Resource | Default | Maximum |
|---|---|---|
| Number of Jupyter servers per user | 1 | 10 |
| Number of CPUs | None | 143872 |
| CPU time | None | 21 d |
| Amount of memory | None | 347 TB |
| Number of GPUs | None | 128 |
| Disk space | None | 6 PB |
| Persistent disk space | None | 185 TB |
Technicals
Some technical insights about PC² JupyterHub:
-
Singularity integration (default + custom container)
-
Remote Desktop with Slurm tools and module
-
Remote Slurm kernel to start only kernels remote on cluster
-
Direct connection of the cluster parallel filesystem