Interconnection networks are crucial to a supercomputer’s performance, acting as its backbone. Without a robust interconnect, even the most powerful systems would struggle to keep up with the demands of modern HPC and AI workloads. By linking supercomputer components like processors and storage with ultra-fast networks, interconnect technologies manage large-scale connections, support parallel tasks, and adapt to data-intensive and AI-driven applications.
NET4EXA builds upon the research carried out in the EuroHPC JU-funded RED-SEA project, which focused on optimising interconnect scalability and performance. Building on the proven success of BXI (BullSequana eXascale Interconnect) versions BXIv1 and BXIv2, NET4EXA will continue the development of the next generation of European Interconnects (BXIv3).
The project aims to enhance these technologies to deliver greater efficiency, capacity, and adaptability for future supercomputing needs. To demonstrate these innovations, the project will deploy a pilot system, ready for integration into exascale and post-exascale infrastructures.
NET4EXA will also look to the future by preparing for BXIv4. By improving interconnect hardware and software, NET4EXA aims to address key challenges such as high-bandwidth, low-latency communication and efficient CPU/GPU integration. Additionally, with a focus on next-generation interconnect solutions, NET4EXA will enable better system performance and energy use for both AI and HPC applications.
More details
The NET4EXA consortium has been selected following the call HORIZON-EUROHPC-JU-2023-INTER-02.
NET4EXA is funded by EuroHPC JU with a total budget of €71 million stemming from Horizon Europe funding programme. The project will run until February 2027.
The project is coordinated by CEA, with a consortium of eight European partners, including leaders from industry and academia.
Background
The EuroHPC JU is a legal and funding entity, created in 2018 and reviewed in 2021 by Council Regulation (EU) 2021/1173 to enable the European Union and participating countries to coordinate efforts and pool resources with the objective of making Europe a world leader in supercomputing.
In order to equip Europe with a world-leading supercomputing infrastructure, the EuroHPC JU has already procured nine supercomputers, located across Europe. Regardless of their location in Europe, European scientists and users from the public sector and industry can benefit from these EuroHPC supercomputers via the EuroHPC Access Calls. This access enables them to advance science and support the development of a wide range of applications with industrial, scientific, and societal relevance for Europe. Additionally, the EuroHPC JU is investing in the development of a European quantum computing infrastructure. This initiative aims to provide access to a wide range of European quantum computing platforms and hybrid classical-quantum systems.
In parallel, the EuroHPC JU funds research and innovation projects to develop a full European supercomputing supply chain: from processors and software to applications to be run on these supercomputers and know-how to develop strong European HPC expertise.
Recently reviewed by means of Council Regulation (EU) 2024/1732, the EuroHPC JU received a new mandate to develop and operate AI factories. These comprehensive open AI ecosystems centered around EuroHPC supercomputing facilities will support the growth of a highly competitive and innovative AI ecosystem in Europe.