Note: this tooling around benchmark testing is very minimal and subject to change!
To run all PD benchmarks, make -C protocol-designer benchmarks
. This will output to a file protocol-designer/benchmarks/output/$(date_time)
. To set the name of the file, specify benchmark_output
These benchmarks use Nanobench, so you can pipe the output to a file and compare benchmark runs with yarn nanobench-compare fileA fileB
- Do not compare benchmarks across different machines.
- Make sure the same resources are available between runs (eg if you kill your dev servers and editor etc, it will likely affect the benchmarks from the run that competed with those processes)