Regensburg 2022 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 44: Poster Session: Statistical Physics and Critical Phenomena

DY 44.15: Poster

Donnerstag, 8. September 2022, 15:00–18:00, P2

Chemfiles: reading and writing atomistic modeling files — •Guillaume Fraux — Institute of Materials, EPFL, Lausanne, Switzerland

Running atomistic simulations produces enormous amounts of data, which has to be post-processed in order to extract scientifically relevant information. Unfortunately, this task is made much harder by the vast menagerie of existing file formats, all containing similar data in a different formatting.

Chemfiles is a software library providing a unified interface to these formats, allowing researchers to spend their time analyzing their data instead of writing file parsers over and over. Chemfiles is implemented in C++, and provides programming interfaces to most of the scientific languages: Python, Fortran, C, Rust and Julia. 21 different formats are currently supported, including both text and binary (i.e. XTC, TNG, DCD, ...) formats. All text formats can be read and written with multiple compression standards (gzip, xz, bzip2). Chemfiles also offers a comprehensive atom selection language, including the unique feature of simultaneous selection of multiple atoms (e.g. pairs: name(#1) == H and name(#2) == O and distance(#1, #2) < 3.0).

Overall, chemfiles one of the fastest libraries for reading files used in atomistic simulation, being between 20% and 10 times faster than other commonly used libraries; while offering a simpler and easy to use programming interface; freeing up time for scientists working with these file formats.