Zentrum für Datenverarbeitung (ZDV) (data center)

bwForCluster BinAC 2

bwForCluster BinAC 2 is part of the implementation concept of the universities of the state of Baden-Württemberg for High Performance Computing (bwHPC) and is available to all scientists in the field of Bioinformatics, Astrophysics, Geosciences, Pharmacy, and Medical Informatics in Baden-Württemberg. BinAC 2 hardware has been optimally tuned to this intended use to achieve high efficiency of applications in these scientific fields.

As headquarters of the nationwide HPC competence center for bioinformatics and astrophysics Baden-Württemberg, this cluster is operated at the Tübingen site by the Zentrum für Datenverarbeitung at the University of Tübingen and the specific applications within the framework of the bwHPC-S5 project are specifically supported.

Hardware

The BinAC 2 system consists of 252 compute and GPU nodes with different hardware configurations:

  • 192 standard compute nodes: 2 × AMD EPYC Milan processors (7543 or 75F3), 64 physical cores per node, 512 GB memory and 450 GiB NVMe SSD local storage.
  • 16 high-memory/SMP nodes: 2 × AMD EPYC Milan processors (7443 or 75F3), 48 or 64 physical cores per node, 2048 GB memory and 14,000 GiB NVMe SSD local storage.
  • 32 GPU nodes: 2 × AMD EPYC Milan 7543 processors, 64 physical cores, 512 GB memory and 450 GiB NVMe SSD local storage. Each node is equipped with 2 × NVIDIA A30 GPUs, each with 24 GB ECC HBM2 memory and NVLink.
  • 8 GPU nodes: 2 × AMD EPYC Milan 7543 processors, 64 physical cores, 512 GB memory and 14,000 GiB NVMe SSD local storage. Each node is equipped with 4 × NVIDIA A100 GPUs, each with 80 GB ECC HBM2e memory.
  • 4 GPU nodes: 2 × AMD EPYC Milan 9555 processors, 128 physical cores, 1536 GB memory and 28,000 GiB NVMe SSD local storage. Each node is equipped with 4 × NVIDIA H200 NVL GPUs, each with 141 GB ECC HBM3e memory and NVLink.

The compute nodes and the parallel file system are connected via a 100 GbE Ethernet network. In addition, 84 standard compute nodes are connected via HDR-100 InfiniBand. Jobs requiring InfiniBand can request these nodes using the --constraint=ib option.

The H200 GPU nodes provide HDR-200 InfiniBand in addition to 100 GbE.

The operating system is Rocky Linux 9.8. Slurm is used as the queuing system, and scientific libraries and software are provided through the Environment Modules system.Module-System.

Learn more:

Access

The computer system is available upon request to all scientists in the fields of bioinformatics, astrophysics, and geosciences in Baden-Württemberg free of charge.

The link for Registeration and Support are given below.

Registration                                                                                                       Support: hpcmaster@uni-tuebingen.de

  1. Share on Facebook (opens in new window)
  2. Share on LinkedIn (opens in new window)
  3. Share on Mastodon (opens in new window)
  4. Share on Bluesky (opens in new window)
  5. Share via email (opens in new window)
  1. Share on Facebook (opens in new window)
  2. Share on LinkedIn (opens in new window)
  3. Share on Mastodon (opens in new window)
  4. Share on Bluesky (opens in new window)
  5. Share via email (opens in new window)