Releasing GPU resources in BOZ analysis

We use the GPU on the online cluster to process the BOZ calibration data. This takes usually ~45min and we do this a couple of time during a beamtime, but the GPU resources stay allocated indefinitely, preventing the next users at the other instrument from using the SA3 online GPU (https://redmine.xfel.eu/issues/221256#change-751556)

We should look into implementing a timeout.

Edited May 26, 2025 by Loïc Le Guyader

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information