Information for Center Selection

All NHR centers have up-to-date HPC hardware and offer best conditions for well-scaling applications. The following is a brief description of the centers' resources as well as the consulting services offered.

Center Description Details Application Guidelines
NHR4CES@RWTH

CLAIX-2018 consists of 1300+ computing nodes, each with 2x Intel Skylake 24-core CPUs and 192 GB main memory. This includes 48 nodes with 2 NVIDIA V100 GPUs each and nodes with up to 2.0 TB main memory. There is a parallel file system (10 PB) and a high-availability file system (1.5 PB). In Q1/2024, a new system with Intel Sapphire Rapids CPUs and NVIDIA H100 GPUs will go into operation.

Support focuses on engineering methods and applications.

Link Link
NHR4CES@TUDa

The compute phases of Lichtenberg II consist of 1217 compute nodes with 2x Intel Cascade-Lake AP (96 cores) or 2x Intel Sapphire Rapids (104 cores) processors and 384-512 GB main memory each, 4x big-memory nodes equipped with 1.5-2.0 TB main memory each, and 11 accelerator nodes with overall 16x Nvidia V100 and 40x Nvidia A100 GPUs. All nodes and the file systems (6 PB) are interconnected via EDR/HDR Infiniband.

The support services focus on methods and applications in engineering sciences.

Link Link
NHR@FAU

NHR@FAU currently runs more than 70 000 CPU cores and offers extensive GPU resources for throughput calculations as well as moderately parallel applications (up to 8 GPUs). Available are 304 NVIDIA A100 (40/80 GB) as well as 352 NVIDIA A40 cards; the latter are particularly suitable for molecular dynamics simulations with standard packages such as GROMACS and AMBER.

User support is known for its expertise in performance engineering for CPU and GPU nodes as well as iterative solvers. In application support, NHR@FAU has a special focus on atomistic simulations.

Link Link
NHR@Göttingen

With the HPC system "Emmy" (111,464 Intel CPU cores), the new partition "Grete" (144 NVIDIA A100 GPUs), as well as modern data management concepts and storage systems, NHR@Göttingen enables running computationally intensive HPC applications, e.g. from the fields of life science, earth system science, fluid mechanics (CFD) and digital humanities, as well as artificial intelligence (AI) and quantum computing applications

The center also offers a comprehensive range of consulting services.

Link Link
NHR@KIT

NHR@KIT operates the "HoreKa" system (60,000 CPU cores, 668 NVIDIA A100 GPUs) for highly parallel (GPU) simulations, and the "Future Technologies Partition" with the latest architectures (ARM CPUs, AMD GPUs, Graphcore accelerators, etc.). There are advanced options for performance and energy measurement.

Dedicated support structures exist for Particle and Astroparticle Physics, Earth System Science, Engineering for Energy and Mobility, Materials Science, Software Sustainability and Performance Engineering.

Link Link
NHR@SW

 

Link  
NHR@TUD


The HPC resources of the NHR@TUD systems provide the scientists with about 94,000 CPU cores and a peak performance of more than 1.5 quadrillion floating-point operations per second. In addition, about 500 NVIDIA V100/A100 GPUs are available for suitable applications. The architecture is specially tailored to data-intensive computations, Big Data analyses and artificial intelligence methods, and offers extensive options for energy measurement and performance monitoring.

NHR@TUD is strongly committed to efficient resource use and offers a wide range of tools, consulting and training services on the above topics.

Link Link
NHR@ZIB

The NHR@ZIB operates the HPC system "Lise" with a CPU partition (121,920 compute cores) for highly parallel applications and a GPU partition (168 Nvidia A100) for data-intensive computing. The wide range of data storage technologies enables individual usage profiles regarding I/O requirements.

Our domain-specific consulting supports research projects from a wide range of applications - e.g. life sciences and earth system sciences - as well as AI as a cross-sectional methodology.

Link Link
PC2

The NHR-center Paderborn Center for Parallel Computing (PC2) is operating the HPC system “Noctua 2” (143.000 CPU-cores, up to 2 TB main memory, 136 NVIDIA A100, 80 FPGAs, https://pc2.de/go/noctua2).

In addition to general support for all areas of science, the highly-qualified scientific support at PC2 focuses on researchers from solid-state physics, material science, chemistry, optoelectronics/quantum photonics/quantum computing, FPGA/GPU development, and machine learning/AI.

Link Link