Separate simple and commit-reveal benchmarking.

[jnaviask]

Currently our benchmarking for the voting module assumes the worst case scenario, which is commit-reveal. However, verifying the commit hash takes significantly more time to perform than calling reveal on a simple proposal, so it might be worth separating the simple and commit-reveal benchmarks to obtain a more accurate computation of the call's weight.

We may also want to establish separate bounds for simple and commit-reveal vote numbers once this is done, factoring in the difference in weight between the two branches.