Make entropy-protocol's integration test into a benchmark

[ameba23]

I attempted to make a benchmark for running the signing protocol (and eventually DKG and proactive refresh) with an arbitrary number of parties in order to inform the subgroups design.

Doing this using criterion worked for me locally but others had issues with it and in the end i abandoned the PR - see #738 (comment)

It would be great to resolve this issue and have a statistical benchmark which we can use to optimise the protocol.

[HCastano]

Will revisit this at some point when we decide to dig deeper into performance metrics.