Guy1m0

@Guy1m0

Joined on Jul 12, 2021

  • Disclaimer: The benchmark settings for each framework have been determined solely based on my interpretation of their respective documentation and codebases. It is possible that misinterpretations have occurred, potentially leading to suboptimal environment configurations, improper testing data preprocessing, or incorrect parameter selection. Consequently, these factors may have influenced the accuracy of the benchmark results. If you detect any such errors or unexpected results, please do not hesitate to contact me via my Telegram account @Guy1m0. I am eager to address any inaccuracies to ensure the benchmarks presented are as reliable and comprehensive as possible. As a hacker who participated in multiple hackathons last year, I often found myself overwhelmed by the challenge of selecting a compelling topic to tackle. Sponsors typically offer several tracks related to zero-knowledge (ZK) with attractive bounties, so how can one quickly develop a ZK-related application to win the bounty within such a limited timeframe? image This question has lingered in my mind since I came second in a ZK Hackathon held by ABCDE Capital last year where I created an anti-cheating ZKML prover system. As a hacker and ZK developer, I often wondered if there are any blogs or resources that could make it easier for newcomers to enter this field and realize the power and intrigue of ZKML. After days of research, I found that awesome zkml is useful but lacks practical real-world applications, making it difficult for hackers to fully understand the capabilities of these frameworks. To address this gap, I have developed a ZKML benchmark as part of the Ethereum Foundation ESP Grant Proposal [FY23-1290]. This benchmark is designed to help developers understand the trade-offs and performance differences among various frameworks. While many frameworks offer their benchmark sets, making direct comparisons is complicated by the numerous variables that affect performance. My approach focuses on creating uniform conditions across all frameworks to provide a practical and straightforward comparison.
     Like  Bookmark