Multi-Kepler GPU vs. multi-intel MIC: A two test case performance study

We present and compare the performances of two many-core architectures: the Nvidia Kepler and the Intel MIC both in a single system and in cluster configuration for the simulation of two physical systems. As a first benchmark we consider the time required to update a single spin of the 3D Heisenberg spin glass model by using the Over-relaxation algorithm. The second application we consider is a reactive fluid-dynamics problem for which we resolve the full Navier-Stokes compressible set of equations without resorting to a turbulence model. The results show that the performances of an Intel MIC change dramatically depending on (apparently) minor details. Another issue is that to obtain a reasonable scalability with the Intel Phi coprocessor in cluster configuration it is necessary to use the so-called offload mode which reduces the performances of the single system. All source codes are provided for inspection and for double-checking the results.