A Max-SMT Superoptimizer for EVM handling Memory and Storage

Abstract Superoptimization is a compilation technique that searches for the optimal sequence of instructions semantically equivalent to given (loop-free) initial sequence. With advent SMT solvers, it has been successfully applied LLVM code (to reduce number instructions) and Ethereum EVM bytecode its gas consumption). Both applications, when proven practical, have left out memory operations thus missed important optimization opportunities. A main challenge superoptimization today handling while remaining scalable. We present $$\textsf {GASOL}^{v2}$$ GASOL v 2 , bytes-size tool smart contracts, leverages previous Max-SMT approach only stack optimize also wrt. storage. can be used size in bytes, aligned with criterion by Solidity compiler consumption. Our experiments on 12,378 blocks from 30 randomly selected real contracts achieve gains 16.42% version optimizer without handling, 3.28% over already optimized .