HPC-Cluster ARA

HPC Cluster Ara

High-performance computing for research projects with parallelised applications
HPC-Cluster ARA
Image: Andre Sternbeck (Universität Jena)

Ara — named after one of the seven wonders of Jena — is a high-performance computing (HPC) cluster that is available to research groups in Computer Science, Physics, Biology and Chemistry for computationally intensive research projects. The HPC cluster is used in particular for massively parallelised calculations, each of which has to be carried out on several computing nodes (high-performance computing).

Ara was funded as part of two DFG large-scale equipment proposals. Additional extensions were made possible by European research funding.

Profile

Compute node of HPC cluster Ara

Image: Andre Sternbeck (Universität Jena)

Ara was put into operation in 2016 and expanded in 2018. The HPC cluster currently comprises 320 computing nodes, some of which have additional GPU accelerators or large working memories.

The system currently comprises the following components:

Compute nodes

Ara has a total of 320 compute nodes with

  • 132 standard compute nodes with 24 cores and 128 GB of RAM per node
  • 168 standard compute nodes with 36 cores and 192 GB RAM per node,
  • 8 HighMem compute nodes with extra-large RAM of 1.0 to 1.5 TB per node and
  • 12 GPU compute nodes with additional GPUs (computing accelerators) of the type NVIDIA P100, V100 and A100.

Storage systems

Ara provides a high-performance parallel file system with a capacity of 520 TB as a "work" partition. In addition, users have a permanent home directory for their own programmes and source codes.

    Network

    All compute nodes and storage systems are connected via a high-performance Omnipath network with a bandwidth of 100 Gbps (Omnipath 100). The network has a spine-leaf topology with a blocking factor of 2:1.

    Miscellaneous

    • The cluster received a new software stack based on AlmaLinux 9 in March 2024.
    • Slurm is used as a workload manager for the distribution of computing tasks to the cluster.

    Note: Ara is only available to research groups involved in the (full) proposals. You can gain access to the resources upon formal request via the URZ service deskExternal link.