Optimise the PCS round in ECCVM

[maramihali]

Currently, on the verifier side, we first execute Shplemini, then reduce the BatchOpeningClaim to a single OpeningClaim by performing the batch_mul delayed by Shplemini. Then, we construct the translation OpeningClaim, and the two are being reduced to a single OpeningClaim by executing a second iteration of Shplonk. Finally, we verify the OpeningClaim via PCS. This could be further optimised as we currently perform 4 batch_muls at the expense of more complex arithmetic