Monte Carlo Simulations on Xeon Phi: Offload and Native Mode

In high performance computing, Monte Carlo methods are widely used to solve problems in various areas of computational physics, finance, mathematics, electrical engineering and many other fields. We have designed Monte Carlo methods to compute Feynman loop integrals in high energy physics, and to solve problems in stochastic geometry with applications to computer graphics, such as the tetrahedron picking problem leading to 12 dimensional integrals.The Intel Xeon Phi is a coprocessor based on a Many Integrated Core (MIC) architecture to gain extreme performance. We have used two different modes, "offload" and "native", to implement the simulations. In offload mode, the main program resides on the host system and the functions are executed on MIC. In native mode, the program is fully executed on MIC.In this thesis, we compare the performance of our applications running on Intel Xeon Phi, in terms of time and speedup, with the sequential execution on the CPU. The comparison results between the different modes are then shown further in the thesis. In addition, the applications are designed in both single and double precision to show the difference with respect to time.