Dynamically generated benchmarks with vbench

Category: scikit-learn
#gsoc #vbench #benchmarking #python

To construct a vbench benchmark you need a setup string and a code string. The constructor’s signature is:

Benchmark(self, code, setup, ncalls=None, repeat=3, cleanup=None, name=None, description=None, start_date=None, logy=False).

Why generate benchmarks dynamically?

For most scikit-learn purposes, the code string will be …