Docs/benchmark postproc

[jamesETsmith]

Summary

This PR adds benchmarks for three operations and compares fast-pauli and qiskit for all three. It also adds some discussion around them on the docs page.

Details

We focus on three operations:

1. Pauli string applied to a state (shows that we can do things quickly)
2. Pauli operators applied to a states (shows that we're still faster even for more complex objects) for a range of different sizes of Pauli operators).
3. Expectation value for a batch of states (shows that we can do expectation values and work with batches or states)

[jamesETsmith]

I'm going to delete the unused figure/benchmarks to reduce clutter and tech debt. When we revisit them, this is a pointer to a commit, where the code exists in commented out form.