Large-scale experiments at CERN and the KM3NeT neutrino telescope are generating far more, and more complex, data than current systems can handle.
To tackle this, research funding agency NWO is investing €3.5 million in MERGE (Machine Learning E‑Infrastructures for Results on the Global Exascale), a new computing infrastructure project led by Nikhef and SURF that will fully exploit artificial intelligence for frontier particle physics research.
The MERGE project will significantly expand the computing capacity available to VU physicists working with LHC and KM3NeT data. VU Amsterdam is closely involved, with our colleague Mara Senghi Soares as one of the co-PIs. MERGE will be using the Snellius supercomputer at Amsterdam Science Park as a central hub. With state-of-the-art graphics processors, accelerator cards and dedicated AI workflows, MERGE makes it possible to process data volumes that will reach exabyte scale once the high-intensity LHC accelerator and KM3NeT are fully operational. According to Nikhef program director David Groep, MERGE will enable a transition from small-scale analysis projects to production-scale operations worldwide, while also improving energy efficiency. Nikhef director Jorgen D’Hondt emphasizes that the new infrastructure is essential to delve deeper into unknown areas of physics. The MERGE facilities are intended to benefit Dutch science as a whole: in principle, other research projects beyond LHC and KM3NeT will also be able to use the upgraded infrastructure.