Welcome to the website of the Scientific Computing service of ZIMT.
ZIMT (Zentrum für Informations- und Medientechnik) operates the OMNI cluster and a number of other systems to provide scientific computation capabilities to all scientists of Uni Siegen. Using the systems is free for all members of the university.
If you would like to use one of our systems, or if you are looking for advice on scientific computing, contact us at hpc-support@uni-siegen.de or visit us in our weekly consultation hour.
Latest News
About this website
This website has several purposes:
- For news updates about the cluster and the other systems
- As a source of information for new users of OMNI, who want to know how to obtain access to the cluster and how to get started once they have access.
- As a source of information for advanced users of OMNI, when they need information about the installed application software and the installed programming languages, libraries and development tools or system information about the cluster.
- As a collection of documentation and best practices.
Systems
OMNI Cluster
The OMNI cluster is the central computing resource and the largest of our systems by far.
The cluster is used for example for computations in the fields of computational fluid dynamics, finite-element methods and chemical simulations. However, it is also designed with applications from many other fields (e.g. Big Data applications like Tensorflow and Pytorch) in mind.
The OMNI cluster consists of 439 regular compute nodes with 64 cores and 256 GB RAM each, as well as a multitude of other nodes for special purposes, including GPU nodes. Around 500 TB of total disk storage is available and the peak performance is roughly 1044 TFlops.
The installed operating system is Rocky Linux.
You can find more detailed system information here.
Contact
The cluster is operated by the HPC support team of ZIMT. If you have questions, problems or would like software to be installed on the cluster, you can reach us at:
or visit us in our weekly consultation hour.
All cluster users are also added to a mailing-list automatically which is used for news updates.