A Performance Study of a Dual Xeon-Phi Cluster for the Forward Modelling of Gravitational Fields

Figure 12

Behaviour of the computation times on the Xeon Dual CPU. One thread is mapped to one core using a compact affinity over the sockets.