LAMMPS' PPPM Long-Range Solver for the Second Generation Xeon Phi

There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

Abstract

Molecular Dynamics is an important tool for computational biologists, chemists, and materials scientists, consuming a sizable amount of supercomputing resources. Many of the investigated systems contain charged particles, which can only be simulated accurately using a long-range solver, such as PPPM. We extend the popular LAMMPS molecular dynamics code with an implementation of PPPM particularly suitable for the second generation Intel Xeon Phi. Our main target is the optimization of computational kernels by means of vectorization, and we observe speedups in these kernels of up to 12x. These improvements carry over to LAMMPS users, with overall speedups ranging between 2-3x, without requiring users to retune input parameters. Furthermore, our optimizations make it easier for users to determine optimal input parameters for attaining top performance.

Related collections

Most cited references 8

Record: found
Abstract: found
Article: not found

Gaussian split Ewald: A fast Ewald mesh method for molecular simulation.

Yibing Shan, John L Klepeis, Michael P. Eastwood … (2005)

Gaussian split Ewald (GSE) is a versatile Ewald mesh method that is fast and accurate when used with both real-space and k-space Poisson solvers. While real-space methods are known to be asymptotically superior to k-space methods in terms of both computational cost and parallelization efficiency, k-space methods such as smooth particle-mesh Ewald (SPME) have thus far remained dominant because they have been more efficient than existing real-space methods for simulations of typical systems in the size range of current practical interest. Real-space GSE, however, is approximately a factor of 2 faster than previously described real-space Ewald methods for the level of force accuracy typically required in biomolecular simulations, and is competitive with leading k-space methods even for systems of moderate size. Alternatively, GSE may be combined with a k-space Poisson solver, providing a conveniently tunable k-space method that performs comparably to SPME. The GSE method follows naturally from a uniform framework that we introduce to concisely describe the differences between existing Ewald mesh methods.