sergei iakovenko
sergei iakovenko
TLDR: let's close it 1. Currently, we are performing a single `batch_mul()` in BN254-based Recursive Verifiers, done in https://github.com/AztecProtocol/aztec-packages/pull/9329 and https://github.com/AztecProtocol/aztec-packages/pull/8351 2. The number of scalar muls is optimal in...
The first part is solved by https://github.com/AztecProtocol/aztec-packages/pull/14144 I don't have any strong opinion re the single proof part, @maramihali did you have any particular reason in mind?
Partly handled by https://github.com/AztecProtocol/aztec-packages/pull/14144 Most of the provers return proof by exporting it from the transcript. Except for the case when `HasIPAAccumulator==true`
@suyash67 and @ledwards2225 seems this can be closed?
To save some work: It can be extracted from my commits related to ZK Sumcheck impl, pls ping me when relevant and I can find it
This transcript pattern also obstructs the sharing of transcripts among several provers. In CIVC `prove()`, we have a sequence of provers MegaZK -> Merge -> ECCVMProver -> Translator, where each...
`ECCVM` num rows for a single msm of size $N$: $33*\lceil N/4 \rceil + 31$ for $N$ short random scalars $33*\lceil N/2 \rceil + 33$ for $N$ ful sized random...
added tests that are confirming the formulas here https://github.com/AztecProtocol/aztec-packages/pull/16262
what's the difference between this and the case of Zm?