NHR@TUD
Technische Universität Dresden
Dresden

The Center for Information Services and High Performance Computing (ZIH) is a department within CIDS - Center for Interdisciplinary Digital Sciences at TUD Dresden University of Technology. The mission of ZIH is to enable excellent interdisciplinary research by providing advanced high-performance computing infrastructure, consulting, and training.
Key Expertise and Profile
- Data-driven HPC and high performance data analytics across key application areas, such as life sciences and earth system science
- Systems are equipped with detailed energy and performance monitoring, enabling in-depth performance and energy-efficiency analyses
- Tiered storage
- I/O performance optimization to support data-intensive and compute-intensive applications effectively.
In the area of scalable data analytics and machine learning we are collaborating closely with the national AI competence center ScaDS.AI Dresden/Leipzig.
Technical Specifications
| Center Name | NHR@TUD |
|---|---|
| Total Resources: | |
| # CPU-Cores | 96 896 |
| # GPUs | 416 |
| Additional Information | |
| max. job run time [h] | 168 |
| jupyter notebook available [y/n] | y |
| CPU Resources: | |
| System 1 | Barnard |
| CPU-Type | Intel Sapphire Rapids 8470 |
| # nodes | 696 |
| # Cores per node | 104 |
| memory per node [GB] | 512 - 1024 |
| System 2 | Romeo |
| CPU-Type | AMD Rome 7702 |
| # nodes | 188 |
| # Cores per node | 128 |
| memory per node [GB] | 512 |
| System 3 | Julia |
| CPU-Type | Intel Cascade Lake 8276M |
| # nodes | 1 |
| # Cores per node | 448 |
| memory per node [GB] | 24576 |
| GPU Resources | |
| System 1 | Capella |
| GPUs per node | 4 x NVIDIA H100 |
| # nodes | 104 |
| memory per GPU [GB] | 94 |