1. 引言
前序博客有:
微软团队2022年论文《SuperNova: Proving universal machine executions without universal circuits》,为Nova方案( 微软团队2021年论文 《Nova: Recursive Zero-Knowledge Arguments from Folding Schemes》)的扩展,可支持:
- recursive proofs of multi-instruction virtual machine executions without paying for universal circuits (i.e., the cost of proving a program step is proportional only to the size of the circuit for the invoked instruction)。
Nova每个step仅支持单一计算,当用于证明machine executions时,需构建 可执行所支持的任意指令的 单一电路,然后Prover将pay for “uninvoked” instructions。
而SuperNova,为一种新的递归证明系统,可:
- incrementally producing succinct proofs of correct execution of programs on a stateful machine with a particular instruction set (e.g., EVM, RISC-V)
SuperNova开源代码实现见: