Implement benchmarks

[ghost]

In the long run, performance is going to matter a lot.

• [ ] Implement microbenchmarks for each operation (prove transfer, prove VPBlind, prove VP, verify proof, verify signature, etc)
• [ ] Implement benchmark for each network role (end user, validator, block proposer, matchmaker, etc)
• [ ] Implement benchmark for blockchain scanning for shielded tx (ref heliaxdev/pbc-playground#6 )
• [ ] Benchmark ark-plonk
• [ ] Benchmark transfer circuit vs separate Spend and Output circuits
• [ ] Implement criterion or other performance regression testing so we can tell when changes improve/degrade performance

[cwgoes]

We have action circuit benchmarks and simple VP circuit benchmarks. We can close this issue when we have benchmarks for all of the atomic operations in Taiga, which should be runnable with some sort of cargo bench command.