Search Research Equipment

High Performance Computing Cluster (HPC) Anunna

Description

With the advent of more powerful and effective scientific instrumentation, the complexity and amount of data being generated is exploding. The ability to process, analyse and ultimately act on this data is becoming a critical element of research success.

In times with emerging AI possibilities and data intensive science, High Performance Computing (HPC) is an indispensable tool. This system gives researchers access to compute power that up till now was hardly available. This leads to a faster time to market, cost reductions, better quality of products and above all the opportunity to explore data and models in a more sophisticated way.

The availability of computing power is of utmost importance and plays a crucial role in the advancement of AI for both research and education, and the impact we, as WUR, aim to make in society.

Set-up HPC Anunna

The HPC Anunna is designed as a high available and versatile platform. Whether your job needs lots of memory, raw compute power, Graphical power for AI workloads or enormous storage demands, this HPC Anunna can handle your requests. The current system contains over 11.000 cores, several GPU nodes with NVidia A100 and AMD MI210 Graphical cards and 3000 TB of parallel storage and a fast internal omnipath network.

As a shared facility the system is used by multiple users. Users insert their jobs in a queueing system and the workload manager software handles these requests in order of priority and available free hardware.

The HPC Annuna is housed in a professional datacenter in a secure environment and with constant temperature and power supply. Different backup solutions are available to safeguard the data from loss. All data in the home-drives are back-up daily. Besides that, users can choose to back-up their compute-data as well.

Technical Details

Hardware

  • 77x Computes: 2x AMD Epyc 7713 2.0GHz/64-core CPU (128 cores total/node), 1024 GB RAM
  • 28 x Computes nodes: 32 cores, 384 GB memory, Intel Xeon Gold, 2.1 GHz
  • 4x Fat nodes: 4 TB memory, 2x AMD Epyc 7643 2.3GHz, 48core, 96 cores/node
  • 2x Fat nodes: 64 cores, 1,5 TB memory, Intel Xeon Gold, 2.1 GHz
  • 4x GPU nodes with per node 4x NVIDIA A100 80GB
  • 2x GPU Nodes with per node 2x AMD MI210 64 GB
  • 3.000 TB Lustre parallel file system (15 GB/s)
  • 100 Gb/s Omnipath network
  • Connected to a fast internet connection (SURFnet)

Software

  • Ubuntu Linux on the nodes
  • Jobs scheduling is done with Slurm
  • SURF SALI software for automatic compute node deployment
  • Compilers: Parallel Studio XE 2017 from Intel, Open64, Gcc
  • Software: R, ASReml, Python, SPARK, Java, Octave, GEOS, Galaxy, MATLAB, Jupyter Hub, OpenOnDemand and many others....
  • File formats: NetCDF, HDF5, GDAL
  • MPI: MPICH, MVAPICH, Open MPI

The software is continuously subject to upgrades and installation of new software packages

Applications

The HPC Anunna can be used for a variety of research applications within the domain of the life sciences and living environment. It can be used for bioinformatics, modelling, simulations, analysis and visualisation of data.

Examples are:

  • AI Workloads
  • For academic, commercial and education purposes (big data analyses)
  • Computational fluid dynamics modelling
  • Weather & climate modelling
  • Soil chemistry modelling
  • Animal & Plant genomic research
  • Population modelling
  • Genome assembly
  • Protein folding
  • Economic policy modelling
  •  Advantages HPC Anunna
  • Fast computing: The HPC Anunna cluster has a computing power of 13 TeraFlops. Because of its size, this HPC Annuna offers users the possibility to compute jobs in parallel and to use multiple nodes for a job.
  • Support and knowledge sharing: support for new and experienced users is available and appreciated by the users. Regularly HPC Anunna-related courses on different levels are organized.
  • Service: the HPC Anunna is a full service platform. Maintenance and technical management are taken care of.
  • Community: the user groups and wiki give users the opportunity to share tips and tricks, experiences, knowledge and developed software. In this way your research can be accelerated.
  • Pay only for what you use: compute power is be charged per used core per second. Storage is charged by the actual usage. Also flat-fee arrangements are possible.
  • Easy start-up: users can easily access and use the HPC Anunna to use their own models. In addition, many software is already installed and ready to use.

Complementary Techniques

WIKI

On wiki.anunna.wur.nl you can find more user information


Last edited by Alexander van Ittersum on 2023-11-08

Our expert(s)
Alexander van Ittersum
Contact our expert(s)
Brand

Dell

Organisation

IT Design
Shared Research Facilities

Subdivision

IT Design



To explore the potential of nature to improve the quality of life.

This Shared Research Facilities equipment database offers researchers from universities, research institutes and companies, including start-ups, the opportunity to find research equipment with their relevant expert and to use it, generally on a pay-per-use basis. The equipment is available at different locations at Wageningen Campus or other (company) locations. If you want to share your own research devices or if you need more information, please contact us.