Just run a test with peak load of 100 TPS of ZK transactions on ETH mainnet. We could do more but hit the limit by consuming the entire capacity of DigitalOcean's largest suitable servers (next time AWS). Block explorer: https://t.co/LV1MA5m789
@btcVeg @ebfull Also looks like naive efficient batching of SONIC proofs on a smart-contract is hardly possible using the smart-contract, cause every proof with advice requires around 13 multiplications in G1 with unique bases per proof
@btcVeg @ebfull Groth16 with 1 input would be 500k, going down to 250-300k in a batched mode. I don’t remember exactly for Pinocchio, but may be around 2m
@ebfull Regarding the transcript - what would you suggest? A ceremony organizer has to verify a contribution and decompress the previous accumulator for a new contribution - that makes good checkpoints - so one can just truncate old records from the transcript assuming it’s verified
@ebfull Code is curve agnostic, uses the same pairing traits as original one, we just formed, extended it and implemented bn254 https://t.co/kOXSuepzli
#ETHSingapore We’ve released a bellman for ETH, a zkSNARK proving system as a part of our work for the real hackathon project, source is here
https://t.co/Ut8y0JSFRg
Practical example is also available, solves important case of Plasma Cash history size
https://t.co/urGZTlwzTP