Benchmark tool for cross comparison of DDA codes.
pip install dda-benchgit clone https://github.com/Argencle/dda-bench.git
cd dda-bench
pip install -e .This package does not ship external DDA solvers. You must provide executables and point your config to valid paths.
Full configuration reference:
This repository is linked to, and was developed as part of, a benchmark paper.
The dataset used for that paper is available in:
To reproduce the equivalent parameter agreement between codes reported in Appendix C of the paper, patches from that data repository must be applied to ADDA and DDSCAT before running dda-bench.
The corresponding benchmark command set is provided in:
dda-bench --initThis creates starter files in current directory:
DDA_commandsdda_codes.jsonbin/
The starter dda_codes.json points to executables under bin/.
--init copies only bin/diel/* and bin/*.dat/bin/*.par support files, not solver executables.
Then run:
dda-benchdda-bench without options expects:
./DDA_commands./dda_codes.json
Override with your own files:
dda-bench --commands /path/to/DDA_commands --code-config /path/to/dda_codes.jsonOther options:
dda-bench --output outputs --omp 1 --cleanThe command writes:
dda_bench.logdda_bench.errors.log- per-case
results.jsonunder output directory summary.csvin output directory
If you use this code in your research, please cite both the software and the paper.
Argentin C. (2026). dda-bench. Zenodo. 10.5281/zenodo.18836855
Argentin, C., Chaumet, P. C., Gross, M., & Yurkin, M. A. (2026).
Floating-point--consistent cross-verification methodology for reproducible and interoperable DDA solvers with fair benchmarking.
arXiv:2603.02871
You can click the “Cite this repository” button on the right-hand side of the GitHub page for citation formats (BibTeX, APA, etc.).