This folder contains files used to parse the results gained during the evaluation.
The scripts here expect a csv input file gained from JUBE.
They can be generated using the -s csv parameter when calling jube result on the finished benchmark folder.
analyse_correctness.py: Using the output JUBE csv of the correctness benchmarks, this creates formatted tables and calculates the accuracy metric.parse_performance.py: Using the output JUBE csv of the performance benchmark, this creates separate csvs per proxy app, which were used to generate the tables in the paper.