A Statistical Analysis of Error in MPI Reduction Operations Dataset

This is the dataset used to generate the figures and information contained in the paper A Statistical Analysis of Error in MPI Reduction Operations by Samuel D. Pollard and Boyana Norris, to appear in IEEE's Fourth International Workshop on Software Correctness for HPC Applications, 2020.

A description is provided in the README.md as well as the software dependencies required to re-generate these data. The log files and tab-separated-values files (tsv) allows a user to analyze the same data we used for the paper. The file datasets-pollard-correctness2020.tar.bz2 decompresses to about 5.7GB.

The source code used to generate these data is available on Github.