is there any way to optimize the MSM of the verifier contract

[cliff0412]

my circuit public input is very big. as my estimation, one public input will lead to one point multiplication and one addition, which is around 6k gas consumption. is there any way to optimize the MSM, e.g. pippenger algorithm or bos coster algorithm. what if the MSM is calculate off chain? people could verify the MSM computation is correct?

[Arvolear]

Hey, @cliff0412, I've tried to rewrite the verifier contracts to use alternative point multiplication algorithms and it turns out that the current approach is the cheapest.

Pippenger and other algorithms usually optimize the number of additions in double-and-add approach. However, Solidity verifiers use precompiled contracts to carry out the job. Substituting these precompile calls with the actual multiplication algorithm degrades the gas cost significantly.

Unfortunately, I haven't been able to find a better approach.

[cliff0412]

thanks for the info

[Arvolear]

One way this may work is if you make current public signals private and pass a hash of these signals as a public one (calculate it in a smart contract). Then you may verify the correctness of the hash within the proof.

Not absolutely sure about legitimacy of this approach, but you may give it a shot :)