| Popis: |
Molecular Dynamics Simulation (MDS) usually requires fast process to do the computation; moreover, the number of particles can be augmented into thousands and million atoms or molecules. Regarding to the great numbers of particles involved the efficiency of MDS algorithm and codes are very important to lead the faster computation time. To achieve such fast algorithm, we can utilize parallel processing/computation to reduce the computation time consume. We report our research on using OpenMP to perform parallel computation in the multicore processor for MDS. It is evaluated the speed up of OpenMP parallel computing to perform the MDS which computes time dependent behaviour of a molecular system, including the position, velocity, and orientation changes in time of each particle. We run the OpenMP based parallel algorithm of MDS for 500, 1000, 2000, 4000, 8000 number of atoms in 1-, 2-, 4-, and 8-threads of 8 cores processer intel i7. The result shows that the speed up of algorithms increases as numbers of thread are used but the more number of threads is used it reduces the effectiveness because of the message passing complexity. At the end of paper it is shown the visualization of simulation run. |
