Baseline measurements

[ntucker]

To make it easier to care across different configurations it would be desirable to be able to run some simple benchmark and then use that as a ratio to the other benchmarks in the suite instead of reporting on actual numbers.

This is especially useful for those running benchmarks automatically in github actions as their servers have quite variability in performance.