Zerocheck optimizations for Irreducible
Zerocheck speedup
About
The problem it solves This project delivers improvements for zerocheck benchmark run at AWS ARM instance for the bounty, described below: Accelerate our benchmark for the zerocheck protocol. The input is 3 multilinear polynomials---a, b, and c---over GF(2) with 24 variables each (2^24 coefficients). The protocol checks that the values a(i) * b(i) - c(i) = 0 for all coordinates i in the boolean hypercube. Variations of sumcheck that send more than 2 elements in the first round are allowed (ie. univariate skip), as long as the first round message is at most 1,024 GF(2^128) elements.