EIPsInsight

Loading Experience

Preparing your insights...

Loading0%
EIPsInsightEIPsInsight
Sign in
All protocol calls
PreviousNext
PQI2026-06-03_042

PQ Interop #042

2026-06-03 1 decision 303 transcript lines

Transcript

Ask about this call

Answered from this call's summary and transcript only.

AI-generated — verify against the recording.

Call summary

Targets
  • •Compile DevNet 4 report with performance metrics before fully closing it out — 00:35:27
  • •lean spec refactor complete for internal EF presentation by end of June — 00:14:13
Decisions
  • •Move primary DevNet testing focus to DevNet 5; Zeam main branch moving to DevNet 5 spec immediately — 00:27:45
Highlights
  • Documentation:
    • ·Eth Research post published: 'Exploring the Design Space for a Post-Quantum Public Key Registry for Ethereum Validators' — covers XMSS, Poseidon, KoalaBear, registry design, open questions; intended as EIP discussion starter — 00:18:55
    • ·lean spec refactor in progress: abbreviations removed, concerns split, folders reorganized, test coverage increased; targeting internal EF presentation by end of month — 00:14:10
  • Client Updates:
    • ·Zeam: DevNet 4 stable for 64 nodes except intermittent snappy P2P encoding error; DevNet 5 interop begun with Ethlambda — 00:08:09
    • ·Qlean: DevNet 5 image ready; interop confirmed with Ethlambda locally; snappy decoding issue under investigation for DevNet 4 — 00:10:30
    • ·Grandine: Updated tag, profiler started, lean spec PRs 781 and 802 implemented; DevNet 5 testing locally — 00:13:54
  • Leanvm Updates:
    • ·LeanVM: Rayon parallelization library being replaced with custom thread pool to stabilize ZK async memory usage and fix architecture-dependent instabilities — 00:17:35
  • Testing Progress:
    • ·Snappy decode error on DevNet 4 gossip is non-deterministic; Lantern suspected as source; all teams asked to run isolated two-client tests — 00:26:03
    • ·Consensus reached to move focus to DevNet 5 rather than wait for full DevNet 4 stabilization — 00:27:45
    • ·DevNet 5 hardware baseline: 8-core/16 GB per node appears sufficient for regular validators; aggregator nodes may need more — 00:43:21
    • ·leanMultisig commit hashes: DevNet 4 = f66d4a9 (~10–15% perf gain); DevNet 5 = 8fcbd77; teams must align on correct commit per devnet — 00:41:31
Action Items
  • •All client teams — Each team run isolated two-client DevNet 4 tests to confirm whether Lantern is sole source of snappy error — 00:32:42
  • •All client teams — Teams align on correct leanMultisig commit hash per devnet (DevNet 4: f66d4a9, DevNet 5: 8fcbd77) — resolve async in group chat — 00:42:16
  • •All teams — Review and comment on Eth Research post on PQ public key registry — 00:24:41

Meeting chat18

  • Thomas Coratger

    https://github.com/leanEthereum/leanSpec/pull/817

  • Ladislaus

    for reference https://x.com/tcoratger/status/2061748291335774427?s=20

  • Thomas Coratger

    https://ethresear.ch/t/exploring-the-design-space-for-a-post-quantum-public-key-registry-for-ethereum-validators/25040/5

  • unnawut

    Ideally before we close out devnet4 completely, I think we should together compile a devnet4 report listing out what we achieved in terms of performance metrics

  • unnawut

    Replying to "Ideally before we close out devnet4 completely, I think we should together compile a devnet4 report listing out what we achieved in terms of performance metrics" so ideally some hours of stable run

  • Gajinder Singh

    Replying to "Ideally before we cl..." once we isolate the issue we can do that, but we will be moving zeam main branch to devnet5 a side branch could continue for devnet4 if needed

  • Mega | ethrex

    Grandine’s (f66d4a974eced803574eb0ea43d812e523c8d7ad) should be fine

  • Thomas Coratger

    https://github.com/leanEthereum/leanSpec/pull/822

  • Thomas Coratger

    8fcbd77958a58666e828315de2d6ce7c93297117

  • Anshal Shukla

    We can take it async

  • Mega | ethrex

    Just in case, that commit is for devnet 5

  • main

    the latest commit is 8fcbd77, I kept the focus on devnet4 to avoid confusion)

  • Parthasarathy Ramanujam

    Those who ran devnet5 image, does it work well on 8 core 16 GB machine?

  • Mega | ethrex

    In our experience, it should run fine

  • Mega | ethrex

    We’re running 16 nodes in a 64GB 8 cores/16 threads machine

  • Parthasarathy Ramanujam

    Reacted to "We’re running 16 nod..." with 👍

  • Kamil Salakhiev

    I ran 64 nodes in 64gb machine for shadow simulation🙂 but in shadow nodes do not compete for CPU

  • Anshal Shukla

    I’ll drop off will continue async thanks!

Key decisions

  • Move primary DevNet testing focus to DevNet 5; Zeam main branch moving to DevNet 5 spec immediately

    DevNet 5Teams decided to move primary DevNet testing focus to DevNet 5 instead of waiting for full DevNet 4 stabilization.

Links