PMDA

Summary

PMDA parallelizes common analysis algorithms in MDAnalysis through a task-based approach with the Dask library. Although still in early development, it is already used on resources ranging from multi-core laptops to XSEDE supercomputers to speed up analysis of molecular dynamics trajectories [1].

Purpose and Capabilities

To speed up analysis of molecular dynamics trajectories, PMDA implements a simple split-apply-combined scheme for parallel trajectory analysis for MDAnalysis [1-3]. The trajectory is partitioned into blocks ("split") and analysis is performed separately and in parallel on each block (“apply”). The results from each block are gathered and combined (“combine”).

PMDA allows one to perform parallel trajectory analysis with pre-defined analysis tasks which have the same input and output as MDAnalysis. In addition, it provides a common interface that makes it easy to create user-defined parallel analysis modules.

Who uses it?

The audience of this package are researchers who are familiar with MDAnalysis and want to speed up analysis of molecular dynamics trajectories. Best results are observed if expensive computations are performed for each frame because I/O (loading each trajectory frame into memory) represents a substantial cost.

PMDA is easy to install and can improve performance of analysis on machines ranging from laptops to supercomputers with hundreds of cores. Like MDAnalysis it runs under Linux, Windows, and macOS and supports both Python 3.5+ and legacy Python 2.7.

References

  1. S. Fan, M. Linke, I. Paraskevakos, R. J. Gowers, M. Gecht, and O. Beckstein. PMDA — parallel molecular dynamics analysis. In C. Calloway, D. Lippa, D. Niederhut, and D. Schupe, editors, Proceedings of the 18th Python in Science Conference (SciPy 2019), Austin, TX, 2019. SciPy.
  2. M. Khoshlessan, I. Paraskevakos, S. Jha, and O. Beckstein. Parallel analysis in MDAnalysis using the Dask parallel computing library. In Katy Huff, David Lippa, Dillon Niederhut, and M. Pacer, editors, Proceedings of the 16th Python in Science Conference, pages 64–72, Austin, TX, 2017. SciPy. doi:10.25080/shinma-7f4c6e7-00a.
  3. Ioannis Paraskevakos, Andre Luckow, Mahzad Khoshlessan, George Chantzialexiou, Thomas E. Cheatham, Oliver Beckstein, Geoffrey C. Fox and Shantenu Jha. Task-parallel Analysis of Molecular Dynamics Trajectories. In 47th International Conference on Parallel Processing (ICPP 2018). doi: 10.1145/3225058.3225128.
← MDAnalysisRADICAL-Cybertools →