Benchmarking suite

[jacobhinkle]

We need a simple way to run benchmarks for our low level functions, in addition to the tests we have which just ensure correctness. What I have in mind is essentially microbenchmarks.

#7 would be a first case where benchmark improvement results would be included when closing the issue.

Relevant links:

• https://perf.readthedocs.io/en/latest/
• https://github.com/ionelmc/pytest-benchmark/
• https://github.com/pytorch/benchmark