An Efficient Optimization of Hll Method for the Second Generation of Intel Xeon Phi Processor

Resumo

In this paper, a new approach to vectorization of algorithms of computational fluid dynamics to simulate the dynamics of astrophysical objects is presented. A co-design of a computational model, from the formulation of equations to software tools, is described. The code performance is analyzed. A speed of 245 gigaflops on Intel Xeon Phi 7250 accelerator and 302 gigaflops on Intel Xeon Phi 7290 accelerator is reached. The code developed is used to solve a problem of interaction of different astrophysical objects such as galaxies, gas clouds, stars clusters.