The European High Performance Computing Joint Undertaking (EuroHPC JU) is a joint initiative pooling together the resources of the EU, 32 European countries & private partners, it has the objective of making Europe a world leader in supercomputing.

The EuroHPC JU is procuring and installing supercomputers across Europe. Wherever in Europe they are located, European scientists & users from the public sector and industry can benefit from these supercomputers. Free access is already being provided to European research organisations, with wider access planned for the future.

In parallel, the EuroHPC JU is funding an ambitious R&I programme to develop a full European supercomputing supply chain: from processors and software to applications to be run on these supercomputers & know-how to develop strong European expertise.

How to cite: Gesenhues, L.: The European High Performance Computing Joint Undertaking (EuroHPC JU) – Leading the way in European Supercomputing, Galileo Conference: Solid Earth and Geohazards in the Exascale Era, Barcelona, Spain, 23–26 May 2023, GC11-solidearth-57, https://doi.org/10.5194/egusphere-gc11-solidearth-57, 2023.

The European Union, together with its Member States, are strongly investing in deploying a world-class, complete and diverse HPC infrastructure, through the recently-created European High-Performance Computing Joint Undertaking (EuroHPC). The consortium formed by Spain, Portugal and Turkey is one of the key actors of this initiative, participating together with EuroHPC in the acquisition and operation of MareNostrum5, that will reach a TCO above 200M€ over 5 years. This system will be accessible through the EuroHPC Extreme Scale calls, as well as the national calls from these three countries. Furthermore, BSC as hosting entity will provide direct user support and open scientific collaborations to research groups interested in using this infrastructure.

How to cite: Pineda, O. and Girona, S.: Access to MareNostrum5 and other European HPC infrastructures, Galileo Conference: Solid Earth and Geohazards in the Exascale Era, Barcelona, Spain, 23–26 May 2023, GC11-solidearth-63, https://doi.org/10.5194/egusphere-gc11-solidearth-63, 2023.

LUMI supercomputer, which is currently #3 on the Top500 list and the most powerful system in Europa, started full production earlier this year. LUMI is jointly funded by EuroHPC JU and a consortium of ten countries led by CSC in Finland.

In this presentation we first discuss the architecture of LUMI from the user’s point of view. More precisely, we introduce the various partitions that make LUMI exceptionally versatile and suitable for a wide array of applications and workflows.

To fully harness the computing power of the system, programmers must be able to utilize the AMD MI250X GPUs of the system. Accordingly, we present the available GPU programming models and paradigms together with the performance analysis tools. We will provide information on the particular strategies to apply based on the initial situation of the application the user wants to be ported and deployed on LUMI; e.g. in terms of existing code-base, programming language, problem size, etc..

Finally, we discuss the access and support model: There are various modes and call for access available from both EuroHPC and the consortium countries. The support is handled by the distributed LUMI User Support Team to which all the consortium countries contribute. The consortium also runs a comprehensive training programme.

How to cite: Zwinger, T., Heikonen, J., and Manninen, P.: LUMI supercomputer for European researchers, Galileo Conference: Solid Earth and Geohazards in the Exascale Era, Barcelona, Spain, 23–26 May 2023, GC11-solidearth-25, https://doi.org/10.5194/egusphere-gc11-solidearth-25, 2023.

The European supercomputer Leonardo is one of the three pre-exascale systems announced by EuroHPC Joint Undertaking and will be a significant step forward in raising European research in the field of computational science.  CINECA has a long history in supplying the most powerful supercomputers in the world. The strong partnership with the EuroHPC initiative has led to the realization of the Leonardo project, a significant step forward in raising European research in the field of computational sciences. 

Leonardo (ranked 4th in the last Top500 list), hosted and managed by CINECA's new data center located in the Technopole of Bologna, will be fully operational within the summer of this year.

Leonardo will have a strategic role in fostering national and international initiatives with a clear focus on Earth Science, actively supporting relevant projects and activities like the second phase (2023-2026) of the EuroHPC Center of Excellence for Exascale in Solid Earth (ChEESE-2P)¹, Geo-INQUIRE², DT-GEO³ and Destination Earth among the others.

In this talk, the Leonardo systems will be introduced, together with the way how CINECA will actively support the scientific ecosystem with emphasis on the support to the Earth community.

[1] funded by HORIZON-EUROHPC-JU-2021-COE-01 under the Grant Agreement No 101093038.

[2] 2022-2024, GA No 101058518

[3] 2022-2025, GA No 101058129

How to cite: Lanucara, P. and Amati, G.: Leonardo: A Simulator4Earth, Galileo Conference: Solid Earth and Geohazards in the Exascale Era, Barcelona, Spain, 23–26 May 2023, GC11-solidearth-55, https://doi.org/10.5194/egusphere-gc11-solidearth-55, 2023.

Accelerated computing is nowadays, de-facto, accepted as the path forward to deploy large-scale energy-efficient scientific and technical computing (including Exascale). The positive side effect has been a tremendous opportunity for domain scientists to accelerate the pace of discovery and innovation, as well being capable to quickly respond and adapt to unforeseen natural scenarios by quickly deploy computational tools in support of coordinated mitigation strategies and on-the-ground responses (so called ‘urgent computing’, work pioneered by the ChEESE Eu Centre of Excellence). The purpose of this talk is to briefly introduce the NVIDIA platform, hardware and software, showcasing few examples of geophysical applications that have been successfully accelerated using NVIDIA GPU and set the stage for the future in computing which involve classic HPC simulations coupled or argument by AI methods.

How to cite: Sarasua, I. and Spiga, F.: Accelerating Time-To-Science in Geophysical Simulations, Galileo Conference: Solid Earth and Geohazards in the Exascale Era, Barcelona, Spain, 23–26 May 2023, GC11-solidearth-59, https://doi.org/10.5194/egusphere-gc11-solidearth-59, 2023.