The cluster bought specifically for the project is now fully operative and ready to perform parallel calculations for the project. It is integrated in the Dahu infrastructure hosted by the GRICAD service in Grenoble.

The cluster consists of 20 nodes each made of two 16-cores Intel Xeon 6130 CPUs at 2.1 GHz (total of 32 cores per node, or 640 cores total). Each node has a total of 192 GB of RAM (6GB/core) and is interconnected by an Intel OmniPath 100 network, allowing massively parallel simulations.

The MHDiscs machine in action.
Categories: Non classé