Supercomputer Lucia

Our latest supercomputer is named "Lucia" in memory of Lucia De Brouckère, chemist and professor at the Faculty of Sciences of the Université Libre de Bruxelles.

The supercomputer is funded by the Walloon Region and is hosted within the A6K ecosystem in Charleroi, and is based on HPE Apollo systems provided by Axians, a subsidiary of VINCI Energies.

Lucia consists of a CPU partition with 300 compute nodes, a GPU partition with 50 compute nodes, and an smaller heterogenous partition for specific workflows such as AI, visualization or large memory requirements, all interconnected via an Infiniband HDR network. Lucia's combined (CPU + GPU partitions) peak performance is about 4 PetaFLOPS, and was ranked 245th on the November 2022 Top500 List with its GPU partition.

The computing infrastructure is connected at high speed (10 Gbps) to the Belgian research network, Belnet, and through it to the European research network GÉANT and the Internet.

You can find a system overview here

The documentation available here has to be completed but you can already find some very useful help.