Dear users,
The maintenance is now over, thanks for your patience.
What was done:
- GPU nodes are now allocated based on their GPU model: base model first and high end latest. It is now possible to request minimum vram per GPU and specify if you want a simple or double precision model. See here for more informations. You’ll see a new column named “weight”. GPUs with low wight will be allocated first.
- Preparation for the home storage quota enforcement. The actual enforcement will be done in the next few days. You’ll receive a notification.
- Batteries replacement for scratch storage server
- Slurm updated to version 22.05.3 (same version as Yggdrasil)
- Kernel updated to latest version
- BeeGFS updated to version 7.2.7
- Mellanox updated to latest LTS version
- All the nodes re installed
- All the servers updated
- Many bug fix (or not so many, as we don’t make mistakes of course:)
Have a nice day, happy computation day 
HPC Team, Yann, Adrien, Gaël, Rémy