Lasso + Jolt: An engineering overview

TL;DR

Lasso and Jolt are two closely related research papers that introduce more efficient lookup arguments for VM design, allowing for faster execution and easier auditability.

Transcript

so next up we've got Michael Shu and Sam Ragsdale from a16z talking about the lasso and Engineering heavy overview so I'll hand you straight over and um enjoy lovely thank you [Applause] hello I'm Sam Ragsdale an investment engineer with a16z crypto and I'm Michael Zoo a research engineer we will be talking about lasso and jolt so these are two clo... Read More

Key Insights

• 👻 Lasso and Jolt provide more efficient lookup argument proving, allowing for faster execution and easier auditability of VMs.
• 👍 Lasso improves the efficiency of proving lookups into massive tables by utilizing chunking operations.
• 🏛️ Jolt builds on top of Lasso and extends it to an entire instruction set, providing a new paradigm for building ZK VMS.
• 🏛️ The developer experience of building a Jolt VM is relatively simple, with easily understandable functions and modular components.
• 👻 By sharing underlying instructions, Jolt allows for the development of multiple high-level VMS using the same toolkit, simplifying the development process and enabling a shared auditing service.
• 🛄 Jolt aims to support existing VMs and provides a roadmap for the implementation of specific instructions and subtables.
• 😑 There are still ongoing developments and optimizations in the implementation of Jolt, including support for pre-compilers and integration with existing VMs.

Questions & Answers

Q: What are the main benefits of using Lasso and Jolt in VM design?

Lasso and Jolt provide faster lookup argument proving and more efficient VM design. Lasso improves the efficiency of proving lookups into massive tables, while Jolt allows for faster execution and easier auditability of the entire instruction set.

Q: How do Lasso and Jolt handle the complexity of large VMs with 64-bit operands?

Lasso provides a solution by utilizing chunking operations, where 64-bit operands are split into eight 8-bit chunks. This significantly reduces the overhead of lookup arguments and makes it more feasible to handle large VMs.

Q: Can Jolt be used with existing VMs or does it require a new implementation?

Jolt can be used with existing VMs but will require the implementation of the individual instructions and subtables. Once implemented, multiple high-level VMS can use the same instructions, providing a shared auditing service and simplifying the development process.

Q: Does Jolt support pre-compilers for VMs that don't have easy pairing free compilers?

While Jolt is still being developed, there are plans to include the concept of gadgets, which would allow for the integration of optimized pre-compile circuits. However, the specific support for pre-compilers will depend on the implementation and future updates of Jolt.

Summary & Key Takeaways

• Lasso is a research paper written by Serena Seti, Justin Thaler, and Riad Wabi that introduces a faster lookup argument prover than existing constructions, significantly improving the efficiency of proving lookups into massive tables.

• Jolt builds on top of Lasso and was written by Erasu Arun, Srinath Seti, and Justin Thaler. It provides a new paradigm for building Zero-Knowledge Virtual Machines (ZK VMS) and applies Lasso to an entire instruction set, allowing for faster execution and easier auditability.

• The presentation discusses the background of lookups and the challenges they present in the context of SNARKs, introduces the concepts of Lasso and Jolt, explores multi-linear extensions, and showcases the developer experience of building a Jolt VM.