GPU-Accelerated Large-Scale Molecular Simulation Toolkit
  • a
  • b
  • c

Coarse-grained molecular dynamics (CGMD) simulations are exceptionally important in the research field of soft matter. In general, CGMD targets problems typically at nano- to meso-scales that are not easily coped with using all-atom molecular dynamics simulations.

Directly inspired and derived from pioneering work of HOOMD-blue developed by Prof. Glotzer鈥檚 group in the University of Michigan, also referred to some other powerful simulation tools such as LAMMPS, GROMACS, MDynamix, and OCCAM, GALAMOST was designed to provide a set of open-source and specific tools of employing NVIDIA GPUs, to accelerate CGMD simulations. These tools are, for example:

1. Modelling 鈥減olymerizations鈥 in a stochastic way in CGMD simulations. The topological connections between beads could be changed according to pre-defined probability. This tool can be used in the studies of chain-growth polymerization, reversible reaction, exchange reaction, and so on.

2. Using anisotropic particle models to describe soft particles as well as rigid ellipsoidal particles with/without patches by combining harmonic repulsive potential/Gay-Berne potential with anisotropic 鈥減atchy鈥 potentials.

3. Using hybrid particle-field technique to accelerate CGMD simulations (G. Milano and T. Kawakatsu, J. Chem. Phys. 130, 214106, 2009). This method could largely speed up some slowly evolving processes in CGMD simulations, such as microphase separation and self-assembly of polymeric systems.

4. Reading in numerical potentials derived from iterative Boltzmann inversion, inverse Monte Carlo, or other structure-based bottom-up coarse-graining methods (Mirzoev et al., Comput. Phys. Commun. 237, 263, 2019), and applying the potentials in CGMD simulations.

In GALAMOST, it is also possible to perform conventional CGMD, Brownian dynamics, and dissipative particle dynamics simulations with various potential forms. The trajectories obtained in GALAMOST could be visualized and analyzed by OVITO.