The increasing complexity of today’s high-performance computing architectures enables accurate and reliable simulations of large-scale problems, as well as the analysis and exploration of large volumes of data.
Moreover, as Artificial Intelligence (AI) continues to evolve with a subsequent increase in data volumes, it has become apparent that HPCs are required to adequately handle large computational workloads. This contributes to accelerating the training, validation and testing of an AI model.
As a result, the University of Luxembourg has set up an online repository, giving researchers the opportunity to demonstrate and discuss the benefits and importance of employing HPC for applications requiring formidable computational power and showcasing real life examples of how these are implemented in practice.
This online repository hosts a number of demonstrators highlighting the capabilities and capacities of HPC in different scientific projects. It aims to show a few research activities currently conducted at the University.
It also gives visibility to the available HPC platforms and informs potential HPC users about the existing HPC expertise of the University research teams.
In particular, the online repository was created to promote the utilisation of HPC in Luxembourg by demonstrating representative case studies and examples from academia with potential applicability in industrial fields.
A call to come
The call for demonstrators was launched at the end of 2020 among research teams of the University of Luxembourg. The HPC demonstrators were selected according to their quality and relevancy to the HPC domain.
A next call will take place within the upcoming months at the University and the online repository will be updated with additional academic success stories.
Applications highlighting technical details, computational complexities, scalability test or other performance evaluation metrics will be highly encouraged from different disciplines and domains. Stay tuned!
In addition of MeluXina
As of 2021, the University HPC facility features two supercomputers (Iris and Aion) totalling a computing capacity of 2,76 PetaFlops and a shared storage capacity of 10,68 PetaBytes.
It complements the national MeluXina supercomputer, which has been specifically designed to meet the needs of the private sector and is among the top 40 of the TOP500 supercomputers worldwide.
MeluXina is a petascale supercomputer, capable of executing up to 18 Petaflops, 18 million billion calculations per second, with 20 PetaBytes of storage capacity and powered by green energy from a cogeneration plant powered by waste wood.
While most HPCs are foreseen as pure research frameworks, 65% of MeluXina’s capacity is available to start-ups, SMEs, large companies and researchers.
Figure: 3D Simulation of X-Rays and γ-Rays propagation in voxelized structures using High Performance Computing. The simulator is able to generate and track 4 billions of photons per second using a dedicated GPU node with four GPU devices (Nvidia V100). Photons’ interactions with matter (photon scattering and absorption) are simulated in the voxelised volume of interest employing Monte Carlo techniques.