Post-Quantum (PQ) Interop #47, July 8, 2026

Meeting Time: Wednesday, July 08, 2026 at 14:00 UTC (60 minutes)

GitHub Issue

Meeting Summary:

This meeting focused on team updates and broker improvements for simulation work across multiple development teams. Parthasarathy reported that the Zim team released a new image with optimizations proposed by ETH Lambda and ETH Lambda, with plans to address remaining spec issues and conduct interrupt testing. Shariq shared that the Rheem team tested various optimizations including skipping type 1 aggregation and 8-second slots, finding stable performance with head and finalization differences of 14-15, though they need alignment on which path to pursue. Mihir from the London team discussed refactoring work to improve code maintainability and resolved interop issues by updating to the latest lean spec commit. Pablo from ETH Lambda reported running interop DevNets with public Grafana metrics, implementing signature aggregation optimizations, and fixing block building issues while temporarily disabling regression aggregation. Boma from Jink Grundin team worked on IT Lambda proposal optimizations and made lean spec fixes. Thomas provided updates on spec work including formal verification fixes and collaboration with the Content2Spec team to extract SSZ specification into an official repo. The team discussed various optimization proposals including skipping compaction, early bird block building, and removing proposer signatures, with current simulations running 16 validators on single servers experiencing CPU contention and aggregation delays due to hardware and server configuration issues.

Click to expand detailed summary

Parthasarathy provided an update on behalf of the Zoom team, reporting that they have been working on suggestions from both the Zoom team and ETH Lambda, including releasing a new image for testing. The team is addressing remaining spec issues and plans to release another updated image later today or early tomorrow, followed by additional interrupt testing. Long-running DevNet testing on servers has not identified any new issues beyond those already reported last week.

Shariq reported on Rheem’s experimentation with various optimizations including skipping type 1 aggregation and 8-second slots, which showed stable performance with head and finalization differences of 14-15. The team is running a DevNet with these optimizations and is sharing the link for review. Shariq also mentioned that Headsner pricing doubled without approval for additional servers, prompting them to seek alternative hosting options, and that the Hive instance needed to be restarted due to full storage with plans to improve data purging.

Mihir provided an update on the Lantern codebase, reporting that the team has been focusing on making it more maintainable through refactorings and testing on a 5-node DevNet setup. They resolved an issue with tracking an old commit for lean spec and identified that images were being built with cache enabled in the Git workflow, which affected performance optimizations. The team disabled the cache and expects to see improvements in aggregations and block building once the updated image is published.

Pablo reported on recent developments including making Grafana public for DevNet metrics, fixing a block building issue, implementing signature aggregation optimizations, and completing Hive spec tests. Boma discussed work on the Itlamba proposal optimization options and lane spec fixes, with local testing in progress and plans to release new images for testing.

Thomas reported on updates to the specification, including small fixes and the implementation of formal verification that helped identify and resolve minor to medium bugs. He mentioned ongoing work with the content specification team to extract SSZ specifications into a dedicated repository, which is currently being tested with Justin Shaglia’s team before being transferred to the official Ethereum GitHub organization. Thomas also noted continued progress on binary fields cryptography, particularly regarding BINU64.

Emil reported progress on LeanVM binary implementation, achieving above 500 XMSs aggregated per second with 10 IFEs. The team is also working on BIMiU64 with Jim to compare with a circuit-based approach. The next step will be to simulate with improvements proposed by Reem, Zim, and East Lambda to observe results.

Shariq explained that they implemented three optimizations including skipping compaction, early bird block building, and removing block proposal signatures from proofs, and are testing these against ETH Lambda on metrics. Mega reported running simulations with 16 validators on two DevNets, each on a single server, and also running a single node DevNet with only lambdas as a smoke test.

The team discussed scaling challenges with validator nodes, where Parthasarathy identified hardware and server configuration issues as root causes of delayed justification and finalization processes occurring in bursts up to 30-45 minutes. Anshal proposed a technical workaround involving two-thirds majority aggregation that could improve performance but noted potential memory allocation concerns if finalization remains delayed. The team agreed to release an official image with the optimization once ready, and decided to continue discussions asynchronously due to travel schedules, with plans to reconvene in two weeks.

Next Steps:

  • Parthasarathy: Address all spec issues raised, release a new image (later today or early tomorrow), and run another round of interrupt testing.
  • Shariq: Align with the team on which optimization path to take forward and push the current optimizations into master if agreed.
  • Shariq: Look into other options for hosting DevNets due to increased Headsner prices and lack of approval for additional servers.
  • Shariq: Investigate better ways to purge old historical data from the Hive instance.
  • Mihir: Continue debugging interop issues, push an update in the Telegram chat once a new image is available.
  • Pablo: Share the link to the public Grafana dashboard in the Telegram channel.
  • Pablo: Share an image of Lambda that can be run with ETH P2P soon.
  • Boma: Release a new image after local testing is done and also release a shadow image for shadow testing.
  • Thomas: Extract SSZ specification from LeanSpec into a specific repo, test it with the Content2Spec team, and transfer it to the official Ethereum GitHub organization once approved.
  • Thomas: Simulate with the two new improvements proposed last week by Reem, Zim, and ETH Lambda, and observe results.
  • Parthasarathy: Release an official image for the optimization discussed (excluding the longest slot) and notify the team once available.

Recording Access:

YouTube recording available: https://youtu.be/uBOrCVnzuwY

YouTube recording available: https://youtu.be/BWcJJq6iAqM