Compute Servers

Compute servers have a lot of memory and CPU cores and can be used interactively to run (graphical) software, scripts, etc. Please note the MaintenanceWindow.

Usage instructions

Interactive compute servers should only be used for applications requiring a GUI and for developing/testing scripts/pipelines. Please use our Slurm batch system whenever possible for your computations

1. Use the getserver -sL command on a Linux computer to connect to the least utilized server.
2. Start your software/script.

Important: All users on a compute server share the same resources without restrictions, so please try to fairly use them. All compute servers are being restarted on the first Wednesday of each month at 7 am.

FAQs

Expand all Collapse all

How can I filter for specific compute servers?

More ...Less ...
The getserver command provides options to filter for compute servers with specific hardware—e.g. a lot of memory, a big tmp directory, a GPU, etc.

Please check the help page for more details:

user@host > getserver --help

Back to FAQ start

Which data directories are available on compute servers?

More ...Less ...
All the study and personal storage is available on compute servers using the same path.

Compute servers have no local storage. Some of them have a big /tmp partition which can be used to temporarily store data.

Back to FAQ start

Which software is available on compute servers?

More ...Less ...
You can use the same software and software environments like on your workstation. It's also possible to run applications requiring a graphical user interface, although they might crash when they need 3D acceleration (OpenGL).

Back to FAQ start