Name: Parallel-analysis-in-the-MDAnalysis-Library
Owner: Becksteinlab
Description: Benchmarking MDAnalysis with Dask (and MPI). Supplementary Information for SciPy 2017 paper.
Created: 2017-02-27 20:18:31.0
Updated: 2017-08-31 20:56:36.0
Pushed: 2017-10-12 00:56:08.0
Homepage: http://conference.scipy.org/proceedings/scipy2017/mahzad_khoslessan.html
Size: 80
Language: Python
GitHub Committers
| User | Most Recent Commit | # Commits |
|---|
Other Committers
| User | Most Recent Commit | # Commits |
|---|
We present a benchmark suite that can be used to evaluate performance for parallel map-reduce type analysis and use it to investigate the performance of MDAnalysis with the Dask library for task-graph based computing (Khoslessan et al, 2017).
A range of commonly used MD file formats (CHARMM/NAMD DCD, Gromacs XTC, Amber NetCDF) and different trajectory sizes are tested on different high-performance computing (HPC) resources. Benchmarks are performed both on a single node and across multiple nodes.
For space reasons, not all data could be shown in the SciPy 2017 conference proceedings paper. For a full analysis see the Technical Report (Khoshlessan and Beckstein, 2017). The report is available on figshare at DOI 10.6084/m9.figshare.4695742.
This repository should be considered part of the Supplementary information to the SciPy 2017 Proceedings paper (Khoslessan et al, 2017).
The repository contain the code to benchmark parallelization of MDAnalysis:
The data files consist of a topology file adk4AKE.psf (in CHARMM PSF format; N = 3341 atoms)
and a trajectory 1ake_007-nowater-core-dt240ps.dcd (DCD format) of length 1.004 ?s with
4187 frames; both are freely available under the CC-BY license from figshare at DOI 10.6084/m9.figshare.5108170
Files in XTC and NetCDF formats are generated from the DCD.
Please raise issues in the issue tracker or ask on the MDAnalysis developer mailing list.
M. Khoshlessan, I. Paraskevakos, S. Jha, and O. Beckstein (2017). Parallel analysis in MDAnalysis using the Dask parallel computing library. In S. Benthall and S. Rostrup, editors, Proceedings of the 16th Python in Science Conference, Austin, TX, 2017. SciPy.
Khoshlessan, Mahzad; Beckstein, Oliver (2017): Parallel analysis in the MDAnalysis Library: Benchmark of Trajectory File Formats. Technical report, Arizona State University, Tempe, AZ, 2017. figshare. doi:10.6084/m9.figshare.4695742