Intern
Center for Artificial Intelligence and Data Science

Computing Resources

General Information

CAIDAS runs a High Performance Computing Cluster combined with a distributed big data storage on the premises of the University of Würzburg. The cluster allows distributed computations on large datasets for research purposes and big data applications. The considerable number of GPUs enable large scale machine learning experiments which are necessary for modern AI research.

Workloads are scheduled in containers through Kubernetes with Ceph providing a distributed filesystem. We also run Hadoop, HBase, and Accumulo within the cluster.

The current cluster also includes an NVIDIA DGX™ A100 system .

 

Technical data

Nodes 22
Physical CPU cores 1432
Logical CPU cores 2864
GPUs 71
System Memory 7900 GB
GPU Memory 994 GB
Total Memory 8834 GB
Theoretical Perfomance (FP32) 950 TFLOPS
Distributed Storage Capacity 1000 TB