Case Study: CERN achieves faster, more energy-efficient HPC simulations with Supermicro BigTwin servers

Thousands of supermicro computers supplied to cern to help advance hpc research

CERN, the European Organization for Nuclear Research, faced the challenge of needing immense and constantly updated computing power to process data from its particle accelerators, like the Large Hadron Collider. The scientists required faster simulation times and increased throughput for their advanced physics research, all while managing the total power consumption of their computing environment. To address this, they turned to their long-standing supplier, Supermicro, and selected the Supermicro BigTwin multi-node servers.

The solution implemented by Supermicro involved the deployment of over 900 BigTwin systems, comprising approximately 3,600 server nodes powered by AMD EPYC processors. This provided CERN with a significant increase in performance, enabling faster and more complex simulations for physics-event reconstruction and data analysis. Supermicro's high-density design also resulted in lower-than-expected energy usage due to shared cooling and power systems, delivering superior performance per watt and advancing the organization's critical research.

CERN

Eric Bonfillou

IT Facility Planning and Procurement