Zstandard Python library benchmark

Fetch the data to run the benchmark on (enwik9 dataset), and prepare all datasets. This will take around 10min to complete.

$ ./prepare_data.sh

Run a quick estimate to limit the time on each test case. For example, the following will make each test be run for up to 10s if needed.

$ ./prepare_benchmark.sh 10

Run the benchmark. This will take a lot of time, depending on the value specified in the last command.

$ ./run_benchmark.sh

Generate the comparison tables as markdown.

$ ./generate_stats.py

Check the results directory.