Add benchmark for the transfer circuit

[moCello]

Describe what you want implemented Create a benchmark that measures the performance of poseidon + plonk with the transfer circuit.

Describe "Why" this is needed We need to see if poseidon + plonk are actually slower than zelbet + plonkup. To do that we will measure their performance with the transfer circuit.

Describe alternatives you've considered N/A

Additional context Even though zelbet is the faster hashing algorithm than poseidon the fact that we need to use plonkup (which is slower than plonk) with zelbet can make the overall performance of poseidon + plonk still better than zelbet + plonkup.

See also zelbet #64