Add more benchmarking documentation

[ernestum]

Adds the benchmarking README and the Benchmark Summary to the Sphinx documentation.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (1af5e4d) 95.67% compared to head (e2a53e0) 95.63%.

❗ Current head e2a53e0 differs from pull request most recent head 6b2c89b. Consider uploading reports for the commit 6b2c89b to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #822      +/-   ##
==========================================
- Coverage   95.67%   95.63%   -0.04%     
==========================================
  Files         102      102              
  Lines        9654     9655       +1     
==========================================
- Hits         9236     9234       -2     
- Misses        418      421       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[qxcv]

LGTM!

(for other reviewers: this is what the link in the README looks like when rendered)

[tomtseng]

using /latest redirects to the most recent release, which I think means we'll need to provide benchmark_runs.zip on every subsequent release — is that something we're committed to doing?

[ernestum]

I would be. What do you think @AdamGleave ? I eventually would love to make this part of the release pipeline.

[ernestum]

Btw I only did type annotation changes in mce_irl.py where codecov indicates changes in coverage. So I think that change is purious and we can ignore it. @AdamGleave would you merge this?