Context: Superoptimization is a synthesis technique that, given a loop-free sequence of instructions, searches for an equivalent sequence that is optimal wrt. an objective function. Superoptimization of Ethereum smart contracts aims at minimizing the size of their bytecode and the gas consumption of executing the contract’s functions. The search for the optimal solution poses huge computational demands –as the search space to find the optimal sequence is exponential on the given size-bound – being the main challenge for superoptimization today to scale up to real, industrial software. Even if the underlying problem for finding the optimal solution is decidable, practical tools often prioritize efficiency over completeness. This means they might be implemented to find a sub-optimal solution or even time out. Objective: This work aims at leveraging superoptimization to a real setting: Ethereum blockchain. This paper proposes a neural-guided superoptimization (NGS) approach which incorporates deep neural networks using (supervised) learning into superoptimization to improve scalability by predicting: (1) if a sequence is already optimal and hence the search can be skipped; (2) the size-bound for the optimal solution in order to reduce the search space. Method: We have downloaded over 13,000 smart contracts deployed on the blockchain for training and testing the machine learning models, and a disjoint set with 100 of the smart contracts with more transactions to prove our scalability gains and impact for the Ethereum community. Results: Incorporating DNNs resulted in a 16x overall speedup (12x for gas) with only 12% optimization loss (14% for gas), or a 3-4x speedup with no optimization loss. For the 100 analyzed contracts, this approach reduced the average compilation time to 3 min per contract and achieved monetary savings of $1.24M. Conclusions: The integration of machine learning models mitigates several limitations of traditional superoptimization by drastically reducing execution times while maintaining most of the original optimization gains.

Program optimization is a key factor for green software. In the context of the Ethereum blockchain, optimization is particularly relevant because there is a fee to pay for each EVM (Ethereum Virtual Machine) instruction executed and also there exist bytecode-size limitations for deploying the software on the blockchain. Still, optimization of EVM code is not as widely spread as one could imagine. This is at least partly due to the lack of trust in the correctness of the tools, as security is even more relevant than efficiency in the blockchain context in which bugs may cause huge economical losses. This article develops a formal verification framework using Coq to ensure the security of EVM optimizations performed on jump-free sequences of EVM bytecode. By means of Coq’s theorem proving capabilities, we are able to automatically verify/certify that an optimized jump-free sequence of EVM opcodes is semantically equivalent to a given original one. We also present an extension to our framework that can handle inter-block optimizations that propagate global information across blocks. We have applied our tool to successfully prove the security of peephole optimizations performed by the standard Solidity compiler, and also to existing EVM superoptimization tools (namely GASOL and Superstack) in which we have found bugs that have been reported and fixed.

Abstract The efficiency and the security of smart contracts are their two fundamental properties, but might come at odds: the use of optimizers to enhance efficiency may introduce bugs and compromise security. Our focus is on (Ethereum Virtual Machine) block-optimizations , which enhance the efficiency of jump-free blocks of opcodes by eliminating, reordering and even changing the original opcodes. We reconcile efficiency and security by providing the verification technology to formally prove the correctness of block-optimizations on smart contracts using the Coq proof assistant. This amounts to the challenging problem of proving semantic equivalence of two blocks of instructions, which is realized by means of three novel Coq components: a symbolic execution engine which can execute an block and produce a symbolic state; a number of simplification lemmas which transform a symbolic state into an equivalent one; and a checker of symbolic states to compare the symbolic states produced for the two blocks under comparison. Artifact: https://doi.org/10.5281/zenodo.7863483