If we use the function bn254_endo_batch_mul_with_generator (which needs to be generalized) we can see very significant savings in the Zeromorph recursive verifier cost, particular in the base rollup VMs where there are about 100 (ECCVM) and 180 (Translator) commitments to batch. This is a 'short scalars optimization'.
If we use the function
bn254_endo_batch_mul_with_generator
(which needs to be generalized) we can see very significant savings in the Zeromorph recursive verifier cost, particular in the base rollup VMs where there are about 100 (ECCVM) and 180 (Translator) commitments to batch. This is a 'short scalars optimization'.