1.1 Goals (Prev) 1.3 Contributors (Next)

Potential Energy

All simpatico programs use a potential energy that can contain:

nonbonded pair interactions
covalent bond (2 body), angle (3 body) and dihedral (4 body) potentials
external one-particle potentials

Several different choices of potential energy function are provided for the nonbonded pair interaction and for each type of covalent interactions. Others are easily added. To simplify integration of MC and MD algorithms, all programs are currently restricted to models with differentiable potentials (i.e., not hard spheres), with no constraints on bond lengths or bond angles. The default version of each program currently includes only non-bonded pair interactions and covalent bond interactions. Other types of interactions can be enabled by setting appropriate flags and recompiling.

Classes that represent particular potential energy functions, such as a Lennard-Jones pair potential, are referred to in simpatico as "interactions". The source files for all interaction classes are in src/simp directory, and all such classes are defined in the "Simp" C++ namespace. The same set of interaction classes is available for use in the mcSim, mdSim, and ddSim programs.

Long-range electrostatic potentials have not yet been implemented in any of the simpatico programs. This is now the most important missing feature.

Molecular Dynamics Integrators

For single-processor MD simulations, mdSim currently provides the following integrators:

NVE velocity Verlet
NVT Nose-Hoover
NPH Anderson barostat
NVT dissipative particle dynamics (DPD)

For parallel domain-decomposition Md simulations, ddSim currently provides:

NVE velocity Verlet
NVT Nose-Hoover thermostate
NPH Anderson barostat
NPT Nose-Hoover thermostat, Anderson / Martina-Tobias-Klein barostat

All of the MD integrators except the mdSim DPD integrator use reversible, symplectic algorithms. The ddSim NPT integrator allows constant pressure simulations to be carried out in cubic, tetragonal or orthorhombic simulation cells, with one, two, or three independent cell dimensions.

Monte Carlo Moves

For single-processor Monte-Carlo simulation, mcSim currently provides the following MC moves:

Single atom displacement
Rigid molecule displacement
Hybrid MC moves (short NVE or NPH simulations as proposed MC moves) in NVT and NPT ensembles.
Reptation move for linear chains
Configuration bias regrowth of all or part of a linear chain.
Double (two-chain) rebridging for linear chains.
Double rebridging moves for ring polymers.
End-swap move for linear heteropolymers (e.g., diblock copolymers)
Semi-grand canonical move (for, e.g., homopolymer blends)

The configuration bias and rebridging move algorithms were all designed for use with flexible bead-spring models of polymers, and have been not modified for use with chemicall realistic models with strong angle potentials.

Data Analysis

Simpatico is distributed with a collection of classes for statistical analysis, referred to as "analyzers". The classes available for use with the single-processor mcSim and mdSim programs include classes that evaluate:

Thermodynamic properties (energy and stress components)
Full and partial radial distribution functions (in coordinate space).
Structure factors (in Fourier space).
Intramolecular correlation functions.
Molecular radii of gyration
Mean-squared distances between specific atoms in a molecule.
Mean-squared displacements for specific atoms or molecular centers-of-mass.
Auto-correlation functions for vector separations of atom pairs within a molecule.
Intermediate scattering (van Hove) function S(k, t) .

All mcSim and mdSim data analysis classes can be used either on-the-fly, during a simulation, or for post-processing of a trajectory file or a sequence of configuration dump files. Postprocessing is done by the the mcSim and mdSim programs, by invoking a command that replaces the main simulation loop by a loop that reads a sequence of configurations from file.

A smaller set of analyzers is distributed for use on-the-fly with ddSim. Classes are available to evaluate thermodynamic properties (energy contributions and stress components) and structure factors. MD trajectories from ddSim simulations can also be post-processed by using mcSim or msSim as analysis tools.

1.1 Goals (Prev) 1 About (Up) 1.3 Contributors (Next)