Accessing Our Services

Attention

The cluster and its services are run and managed by the participating institutions. Some scientific staff are doing this mostly part-time next to their main job as researchers. All hardware was procured by the institutions (some of it via joined funding).

We do provide some amount of resources to non-participating chairs with the intention for people to try out our infrastructure/services before deciding whether to participate. For more information please take a look at our documentation on Participating and Access Levels.

Offered Services

We provide computing resources for GPU and CPU workloads as well as storage for large amounts of data.

Computing Resources

There are two options for computing resources:

  1. JupyterHub: a Jupyterlab server running on our servers. Easy to start with and well-suited for shorter experiments.

  2. Compute containers: a container running on our servers, where you have (almost) total control over your environment, but takes more time to setup.

We preferably give access to Jupyterhub. If you need a container, please explain why you need it in your resource request. For special purposes, we give access to Virtual Machines.

For more details on the available hardware, see the server page.

Big Data Storage

General Purpose Big Data Storage is handled via the S3-compliant Object Store Min.io.

LLM Inference Access Key

An inference service for the LLM model is available. For more information, see the LLM Inference Service. This service is only available for users with an access key. If you are not yet a registered user, request access to the the inference key manager from an administrator. Refer to the section below from more details on how to request access. Otherwise, you have access to it already and can request an access key from the UI at this url.

Resource Access

Please fill out the following form, which will support and generate the request for you (JavaScript required).

Should you not have an email client installed on your machine, please send the preview of the request to resource-request@innkube.fim.uni-passau.de.