Lurk 0.5 Benchmarks | Argument Computer Corporation
Skip to the graphs, the full results and methodology, or read on to discover how Lurk has evolved to empower developers with unparalleled performance and usability in zero-knowledge programming.
Just over three years ago, we set out to create Lurk with the goal of making it the best zk programming language — a tool that would make it easier for developers to leverage the power of zero-knowledge proofs in their applications.
The first version of Lurk laid a strong foundation in terms of expressiveness, ease of use, and data privacy. However, we knew that performance was an area where we would face intense competition from other zkVM projects.
Our bet on Nova, though it was a very new technology at the time, was driven by our vision to bring the highest-performing zkVM to our users. This led us to develop Arecibo, our first observatory prover backend, which played a key role in polishing an early version of SuperNova and bringing NIVC online. Alongside ongoing optimization efforts, these initiatives allowed us to make good progress with our Lurk Beta release.
However, benchmarks remained a crucial but elusive measure of our progress. The Fibonacci benchmark, though not entirely representative of Lurk’s real-world usage, emerged as an industry-standard for measuring zkVM performance. While it posed challenges, it also fueled our creativity and led us to develop new solutions.
