โ Scribed by Zbigniew A. Rycerz
No coin nor oath required. For personal study only.
A method for removing data dependency in an indirectly indexed DO loop is presented. Vectorization of such DO loops may, in some applications and on those computers which cannot do it themselves, result in a considerable speed-up of the calculations. Timings for scalar/vector calculations have been performed on an ETA lOP supercomputer and the application of the algorithm for a speed-up of molecular dynamics (MD) simulations of order N, handling short-range interactions, is discussed. The algorithm presented does not use any special calls and, therefore, it can be applied on most vector computers. The method makes it possible to execute an MD program of order N with practically the same efficiency on any supercomputer no matter if such a loop can be automatically vectorized or not.
๐ SIMILAR VOLUMES
The algorithm is based on a grid search and on partitioning of the atoms into boxes rather than on calculations of distances. The effort associated with the generation of the box list grows only linearly with the number of atoms. The algorithm is particularly advantageous for solvated systems. Test
computer. The force is calculated with sufficient accuracy for practical MD simulations. The processor also calculates virials simultaneously with forces for use in the calculation of pressure, accommodates periodic boundary conditions, and can be used in Ewald summations. An MD Engine system consis
## Abstract **Summary:** The structure of polymer brushes is investigated by dissipative particle dynamics (DPD) simulations that include explicit solvent particles. With an appropriate choice of the DPD interaction parameters $a\_{ij}$, we obtain good agreement with previous molecular dynamics (MD