Using the HPC cluster

Access

In order to use the Lichtenberg high-performance computer for scientific purposes, two basic requirements must be met: a project has to be set up to which the computing time used can be booked. And there has to be a valid user account and associated with a project.

Access to the cluster for users is only possible via the login nodes – a current list of available login nodes can be found here .

All important information about the login nodes, public key, ciphers, MACs, Kex Algorithms and Host Key Algorithms can be found here .

Computing in the cluster is organised via the Slurm batch system. Slurm calculates when and on which of the compute nodes the job is started. Among other things, the demand for resources, the waiting time of the job and the priority of the assigned project are taken into account.

More about “Working with SLURM”

By default, jobs are executed on the entire cluster and thus on nodes of all types and phases (expansion stages).

In order for a job to be executed on a specific or special hardware (equipment), the corresponding resources must be requested.

In general, the distinction is by processor type and by accelerator type, but can also selected by sections.

Learn more"

Various file systems are available for user data.

Learn more

Use of the Lichtenberg HPC is only possible via projects that have been allocated the approved number of core hours.

Learn more

Tips & tricks

Helpful information

FAQ

Important answers to questions about access