pyJac: analytical Jacobian generator for chemical kinetics

This repository contains the source for our paper describing an analytical Jacobian generator for chemical kinetic models. It makes use of the pyJac package, which has been developed concurrently. The paper was published in Computer Physics Communications:

Kyle E. Niemeyer, Nicholas J. Curtis, and Chih-Jen Sung. 2017. "pyJac: analytical Jacobian generator for chemical kinetics." Computer Physics Communications, 215:188–203. https://doi.org/10.1016/j.cpc.2017.02.004

and a preprint is available at arXiv:1605.03262 [physics.comp-ph].

All of the data, plotting scripts, and figures associated with the paper can be found on Figshare:

Niemeyer, Kyle; Curtis, Nick; Sung, Chih-Jen (2017): Data, plotting scripts, and figures for "pyJac: analytical Jacobian generator for chemical kinetics". figshare. https://doi.org/10.6084/m9.figshare.4578010.v1

To see a current build of the paper from the master branch of this repository, refer to https://niemeyer-research-group.github.io/pyJac-paper/ (powered by gh-publisher and inspired by the multiband_LS repository).

Feel free to submit comments or feedback via the Issues tab on this repository.

Reproducing the Paper

The LaTeX source of the paper is in the top directory.

To reproduce all of the figures in the paper, first install packages from the standard Python scientific stack: numpy, scipy, and matplotlib. Then, from the top directory, the five figures in the paper can be generated using our data by:

$ python plotting_scripts/plot_cpu_comparison.py
$ python plotting_scripts/plot_gpu_scaling.py
$ python plotting_scripts/plot_gpu_comparison.py
$ python plotting_scripts/plot_cpu_scaling.py
$ python plotting_scripts/plot_ch4_pasr_data.py

The underlying data can be reproduced by installing the pyjac package, available multiple ways:

1. The easiest way is to install via conda:

$ conda install -c slackha pyjac

2. You can also install using pip:

$ pip install pyjac

3. If neither of the previous methods are available, you can also download the source code from GitHub (https://github.com/SLACKHA/pyJac) and install using setuptools:

$ wget https://github.com/SLACKHA/pyJac/archive/master.zip
$ unzip master.zip
$ cd pyJac-master
$ python setup.py install

Then, all of the functional and performance test results can be reproduced (albeit on different systems for the latter, which will alter values) by using the model and PaSR input files given in the data repository mentioned above (https://doi.org/10.6084/m9.figshare.4578010.v1).

Demonstrating TChem's Lack of Thread-Safety

The folder tchem_multithread_test/ holds a self-contained example demonstrating TChem's lack of thread safety. See tchem_multithread_test/README.md for more details.

License

This work is licensed under a Creative Commons Attribution 4.0 International License. See the LICENSE.txt file or follow the link for details.