Las Palmeras Molecular Dynamics: A flexible and modular molecular dynamics code

详细信息    查看全文

• 作者：Sergio Davis ; Claudia Loyola ; Felipe Gonzá ; lez ; Joaquí ; n Peralta
• 关键词：Computer simulation ; Molecular dynamics ; Plug-ins ; Visualization
• 刊名：Computer Physics Communications
• 出版年：2010
• 出版时间：December 2010
• 年：2010
• 卷：181
• 期：12
• 页码：2126-2139
• 全文大小：1162 K

文摘

Las Palmeras Molecular Dynamics (LPMD) is a highly modular and extensible molecular dynamics (MD) code using interatomic potential functions. LPMD is able to perform equilibrium MD simulations of bulk crystalline solids, amorphous solids and liquids, as well as non-equilibrium MD (NEMD) simulations such as shock wave propagation, projectile impacts, cluster collisions, shearing, deformation under load, heat conduction, heterogeneous melting, among others, which involve unusual MD features like non-moving atoms and walls, unstoppable atoms with constant-velocity, and external forces like electric fields. LPMD is written in C++ as a compromise between efficiency and clarity of design, and its architecture is based on separate components or plug-ins, implemented as modules which are loaded on demand at runtime. The advantage of this architecture is the ability to completely link together the desired components involved in the simulation in different ways at runtime, using a user-friendly control file language which describes the simulation work-flow.

As an added bonus, the plug-in API (Application Programming Interface) makes it possible to use the LPMD components to analyze data coming from other simulation packages, convert between input file formats, apply different transformations to saved MD atomic trajectories, and visualize dynamical processes either in real-time or as a post-processing step.

Individual components, such as a new potential function, a new integrator, a new file format, new properties to calculate, new real-time visualizers, and even a new algorithm for handling neighbor lists can be easily coded, compiled and tested within LPMD by virtue of its object-oriented API, without the need to modify the rest of the code.

LPMD includes already several pair potential functions such as Lennard-Jones, Morse, Buckingham, MCY and the harmonic potential, as well as embedded-atom model (EAM) functions such as the Sutton–Chen and Gupta potentials. Integrators to choose include Euler (if only for demonstration purposes), Verlet and Velocity Verlet, Leapfrog and Beeman, among others. Electrostatic forces are treated as another potential function, by default using the plug-in implementing the Ewald summation method.

ss=""h4"">Program summary

Program title: LPMD

Catalogue identifier: AEHG_v1_0

Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEHG_v1_0.html

Program obtainable from: CPC Program Library, Queen's University, Belfast, N. Ireland

Licensing provisions: GNU General Public License version 3

No. of lines in distributed program, including test data, etc.: 509 490

No. of bytes in distributed program, including test data, etc.: 6 814 754

Distribution format: tar.gz

Programming language: C++

Computer: 32-bit and 64-bit workstation

Operating system: UNIX

RAM: Minimum 1024 bytes

Classification: 7.7

External routines: zlib, OpenGL

Nature of problem: Study of Statistical Mechanics and Thermodynamics of condensed matter systems, as well as kinetics of non-equilibrium processes in the same systems.

Solution method: Equilibrium and non-equilibrium molecular dynamics method, Monte Carlo methods.

Restrictions: Rigid molecules are not supported. Polarizable atoms and chemical bonds (proteins) either.

Unusual features: The program is able to change the temperature of the simulation cell, the pressure, cut regions of the cell, color the atoms by properties, even during the simulation. It is also possible to fix the positions and/or velocity of groups of atoms. Visualization of atoms and some physical properties during the simulation.

Additional comments: The program does not only perform molecular dynamics and Monte Carlo simulations, it is also able to filter and manipulate atomic configurations, read and write different file formats, convert between them, evaluate different structural and dynamical properties.

Running time: 50 seconds on a 1000-step simulation of 4000 argon atoms, running on a single 2.67 GHz Intel processor.