There are 2 distinct Fractal servers for UZH:
Fractal Science Cluster server
The Science Cluster server has access to the scalable compute power of the UZH Science Cluster and exclusively uses the Science Cluster’s scalable storage system for data storage. It’s most suited to large workflows (either large datasets or heavy compute requirements like using multiple GPUs). The Science Cloud server uses distinct virtual machines for each research groups that can have direct access to their storage system (e.g. access to ZMB storage, Central IT shares or custom NAS shares). It is most suited to workflows where data transfer to the cluster is a limiting factor or cluster overheads dominate workflow runtimes (typically when processing sub 100 GB datasets in workflows that aren’t GPU heavy).