gcubed.benchmarking

gcubed.benchmarking contains the functionality required to generate and store all the information required to determine if the results from running one G-Cubed model match the results from running another G-Cubed model.

The benchmarking system is designed to enable a set of benchmarks to be instantiated, for comparison purposes, using any version of the G-Cubed Python package. Thus, benchmarking can be used, and indeed is intended to be used to compare the results from running different versions of the G-Cubed Python implementation on the same G-Cubed model.

Another use case is to compare the benchmark results from running two different models using the same G-Cubed Python implementation. This will enable deeper analysis of the causes of differences in simulation results across the two different models.