Parallel processing and visualization for results of molecular simulation problems

In this paper authors presents “mmdlab” library for the interpreted programming language Python. This library allows to carry out reading, processing and visualization of the results of numerical calculations in the tasks of molecular simulation. Considering the large volume of data obtained from such simulations, there is a need in parallel realization of algorithms for processing those volumes. Parallel processing should be performed on multicore systems, such as common scientific workstation, and on super-computer systems and clusters, where the MD simulations were held. During the development process we have study the effectiveness of the Python language for such tasks, and we have examined the tools for it’s acceleration. As well, we studied multiprocessing capabilities and tools for cluster computation using this language. Also we have investigated the problems of receiving and processing the data, located on multiple computational nodes. This was prompted by the need to process the data, produced by parallel algorithm, that was executed on multiple computational nodes, and saves its output on each of them. As a tool for scientific visualization was chosen an open-source “Mayavi2” package. The developed ”mmdlab” library was used in the analysis of the results of MD simulation of the gas and metal plate interaction. As a result, we managed to observe the effect of adsorption in details, which is important for many practical applications.

parallel processing, visualization, molecular dynamics, Python, Mayavi2

1. V.O. Podryga, S.V. Polyakov, D.V. Puzyrkov, “Supercomputer Molecular Modeling of Thermodynamic Equilibrium in Gas-Metal Microsystems” (in Russian), in Vychislitel'nye Metody i Programmirovanie [Numerical Methods and Programming], vol. 16, no. 1, pp. 123-138, 2015 (in Russian).

2. Python official documentation. 04, Feb. 2016, https://www.python.org/

3. P. Fernando, E.G. Brian, “IPython: A System for Interactive Scientific Computing” (in English), in Computing in Science and Engineering, vol. 9, no. 3, pp. 21-29, 2007. (2015, Feb. 4), [Online]. Available: http://ipython.org

4. Stéfan van der Walt, S. Chris Colbert and Gaël Varoquaux. The NumPy Array: A Structure for Efficient Numerical Computation, Computing in Science & Engineering, 13, 22-30 (2011), DOI:10.1109/MCSE.2011.37

5. Numba official documentation, 04, Feb. 2016, http://www.numba.pydata.org/

6. Vanovschi V., Parallel Python Software, http://www.parallelpython.com

7. Ramachandran, P. and Varoquaux, G., `Mayavi: 3D Visualization of Scientific Data` IEEE Computing in Science & Engineering, 13 (2), pp. 40-51 (2011)

8. John D. Hunter. Matplotlib: A 2D Graphics Environment, Computing in Science & Engineering, 9, 90-95 (2007), DOI:10.1109/MCSE.2007.55

9. Paramiko official documentation, 04, Feb. 2016, http://www.paramiko.org/

10. ImageMagick official documentation , 04, Feb. 2016, http://www.imagemagick.org/