How to use...

In this section, users can find basic information and ‘How-to’ guides for the GWDG HPC systems.

We recommend to start with Slurm, the batch system that gives you access to the HPC resources in various compute partitions. If you are building your own software from source code, you will find the description of pre-installed compilers and examples useful. Usually, the software itself is installed via Spack already and made available in the form of modules. To get the best performance, an informed choice about the most suitable storage system is essential. You will also find instructions on how to handle data transfers and use terminal multiplexers for long-running interactive sessions. Last but not least the section about SSH explains how to configure your own client to access the HPC systems remotely.

• Slurm
  • Multiple programs and multiple data
  • GPU Usage
  • Interactive Jobs
  • Job arrays
  • Job runtimes and QoS
  • Multiple concurrent programs on a single node
  • Resource monitoring and Reports
  • Self-Resubmitting jobs
• Compute Partitions
  • CPU Partitions
  • GPU Partitions
• Spack
  • Spack FAQs
• Storage Systems
  • Data Stores
  • Cluster Storage Map
  • Quota
  • Temporary Storage
  • Project Map
  • Optimizing IO Performance
• Data Transfers
• Data Sharing
• Terminal multiplexer
• SSH

Services

Some of our services and software packages have dedicated how-to pages:

• Chat AI
• Jupyterhub
• Apptainer
• Voice AI