Beyond standard computing frames (mail, web, accounting systems …) CERFACS’ research works rely on availability and intensive use of High Performance Computing.
- computers allowing to run the numerical simulations developed by the teams,
- high storage capacities housing the simulations’ results,
- pre and post-processing services that prepare and ease the analysis of obtained results,
- fast interconnection networks for both internal and external exchanges between the different computing and storage resources.
Today, CERFACS hosts 2 clusters for a global peak of about 1.3 Pflops/s, a primary space disk of 3.4 Petabytes inside the clusters and a secondary space disk of 3 Petabytes hosted on a Spectrum Scale server.
All these servers are interconnected on a backbone at 10 Gb/sec. Accesses to our partners’ computers (Météo-France and CCRT), and to national resources for research (Genci) and international ones (Prace and Incite) enlarge our global simulation capacities.
Head of I.T. Management :
HPC software engineer :
System engineers :