Create a virtualenv (python3 -m venv venv) and activate it. Install requirements with pip install -r requirements.txt. Add benchmarks as python functions prefixed with test_ and have the function take in the benchmark fixture. An example is in numpy/windowing.py.
To run the benchmarks, use make python-bench to run all python benchmarks. Use the BENCHES flag to choose a particular set of benchmarks to run.
To write C++ benchmarks, we use the google/benchmark repository. Look at the example of taco/windowing.cpp for an example benchmark. Run make taco-bench to compile the taco benchmarks.