All Core Devs - Execution (ACDE) #211, May 8

[timbeiko]

All Core Devs - Execution (ACDE) 211

• May 8, 2025, 14:00-15:30 UTC
• Stream
• Ethereum Protocol Calls Calendar subscription

Agenda

• Pectra Mainnet Deployment
• Fusaka
  • PeerDAS → ACDT
  • SFI'ing EIP-7935: Set default gas limit to XX0M
    • RLP Execution Block Limit
    • EIP-7938
• Removal of dynamic jumps in EVM
• State Growth Metrics
  • https://hackmd.io/ob66NuW9QBORHLz7emhSCw?both
  • https://hackmd.io/D12VBHdMSU6y_vcqWcJV_g?both
• Community Consensus, Fork Headliners & ACD Working Groups
• https://ethereum-magicians.org/t/for-an-acd-platform/24098
• EIP-7636 feedback

🤖 config

• Duration in minutes : 90
• Recurring meeting : true
• Call series : ACDE
• Occurrence rate : bi-weekly
• Already a Zoom meeting ID : false #
• Already on Ethereum Calendar : false #
• Need YouTube stream links : true #
• Facilitator email: tim@ethereum.org
  Note: The zoom link will be sent to the facilitator via email

[github-actions]

Discourse Topic ID: 23835

• Action: Updated at 2025-05-08 14:08 UTC
• URL: https://ethereum-magicians.org/t/23835

Zoom Meeting: Reusing meeting 88269836469 for series 'acde'.

• Zoom link not available (Link not found in mapping).

Existing YouTube Stream Links (Reused):

• Stream 1 (May 08, 2025): https://youtube.com/watch?v=Y7j8FdRkSrA

Calendar Event Updated

• Google Calendar
• Skipping email notification (already processed issue All Core Devs - Execution (ACDE) #211, May 8 #1500 before).

Telegram Notification

• Updated message for this occurrence to Ethereum Protocol Updates channel

[MicahZoltu]

I don't think ethereum/EIPs#9678 is ready to be discussed on ACD yet. It doesn't suggest any actual change, and its discussion link is likely to change per editor comments (I think a new thread will be necessary for this). This means there is nothing to be talked about yet, and there is not really an appropriate place to gather thoughts and data prior to the call.

As an example, the arguments for/against an increase of the default gas limit to 36M are wildly different than arguments for/against an increase to 150M, but this EIP doesn't give any hint/indication which of these changes are being proposed here. Without that, a proper semi-permanent discussions-to link, and ideally at least a draft EIP I think an ACD discussion will be fairly non-productive as it will just be debating against strawmen without having any time to prepare arguments/data or analyze people's arguments with contrary data.

[will-corcoran]

Please note today was the final PeerDAS Breakout Room. Moving forward PeerDAS will be covered on the Monday testing calls.

[SirSpudlington]

Is it possible to add the adoption of EIP-7636 to the agenda? Because a while ago I submitted a PR to the go-ethereum repo to add support, however, it was rejected by one of the maintainers due to concerns about the EIP itself. I am debating whether the EIP should be withdrawn even though it has support in Lighthouse and Grandine, and would value the feedback of some of the core devs.

Edit: From the current DHT state, I found the following data points:
The average increase of the ENR size for EIP-7636 supporting clients is 18 bytes (8.86%)
The average size of ENR is 140 bytes
The maximum seen size of the ENR is 250 bytes

I may not be able to attend the call today (but I will be watching the stream after), so if under time pressure you should skip over this topic.

[timbeiko]

@MicahZoltu we've discussed (and agreed) on the last two ACDEs to "include" and EIP in Fusaka that mentions raising the default gas limit. It's unusual, I agree, but we went back and forth on the approach and landed on this. The expectation is that by the fork, we know what value is safe and move defaults towards that.

@will-corcoran ty, noted!

@SirSpudlington if there's time, we will cover it, but it may be easier to get an answer to this async.

[timbeiko]

I've drafted a more specific follow up to my previous "Reconfiguring AllCoreDevs" post, including some reflections on the process surrounding EOF: Community Consensus, Fork Headliners & ACD Working Groups

It's a last minute addition to the agenda, and quite long, so I don't expect people to have reviewed it before the call. If time permits, happy to give an overview or to leave space for others to voice their opinions on the path forward.

[gcolvin]

At this point I'd like to look for consensus on whether people want to remove dynamic jumps from the EVM at all. Ever.

[MicahZoltu]

@MicahZoltu we've discussed (and agreed) on the last two ACDEs to "include" and EIP in Fusaka that mentions raising the default gas limit. It's unusual, I agree, but we went back and forth on the approach and landed on this. The expectation is that by the fork, we know what value is safe and move defaults towards that.

The problem I have here isn't with gas limit default changes or having an EIP to describe that, it is that the EIP that is slated to be discussed doesn't actually have any content in it that allows for any discussion prior to ACD. Since EIP-7935 doesn't actually suggest any change, it is not possible to formulate an argument against it nor gather data to argue against it in the upcoming ACD.

I have been desperately looking for an appropriate place to air my concerns with gas limit increases, but the only place that was even close was Vitalik's recent Ethereum Magicians post which is not a topic of discussion for the upcoming ACD call. I tried discussing in the gas-limit-testing channel in Discord and was asked to leave. Dankrad's EIP at least was well formed and gave a concrete starting point for discussion and I was able to formulate some arguments against it, but that also isn't up for debate in the upcoming call.

I feel like before any change proposal is brought to the ACD call, there should be place to discuss the merits of that change proposal prior to the call both to minimize time waste on the call with discussions that could have been had on a forum or chat, and to give participants time to properly prepare for a discussion. At best, EIP-7935 just lets people know that someone will be proposing something related to gas limit changes on the call and one will have to wait for the call to find out what that proposal is.

---

For a more specific concern around the proposal that I haven't seen any discussion on at all (except for mine) is how gas limit increases will impact RPC providing nodes. RPC providing nodes get no/minimal benefits from DAS, Statelessness, history expiry, Portal Network (at the moment), etc., which are all of the things people cite when they say "higher gas limit is fine because ...". All discussion is centered around "how much gas can we jam into a block before staking nodes stop working", which IMO isn't particularly relevant to the actual reason why we have low gas limits.

To quote a client core dev that I would prefer not to doxx (they can claim credit if they like):

RPC is the biggest handwaving I see from Ethereum leadership when they talk about the roadmap

[MicahZoltu]

Also it is worth noting that the link to https://eips.ethereum.org/eip-7935 doesn't show a discussions URL. Could just be a bug in the templating engine that needs fixed, but also makes it hard to engage prior to the call (I can work around this myself I think). Fixed in ethereum/EIPs#9730

[dankrad]

I would like to discuss EIP-7938

I think we need a long term predictable L1 scaling schedule.

[Marchhill]

Related to the discussion on EOF and ACD improvements, I'd like to briefly suggest making a new platform for ACD. I made an argument for it here

[CPerezz]

I'd like to mention a small call for participation on state-growth related metrics obtention.

https://hackmd.io/ob66NuW9QBORHLz7emhSCw?both (Data points that people want to collect)
https://hackmd.io/D12VBHdMSU6y_vcqWcJV_g?both (State bloat attacks or state-growth scenarios that people want to try running to maximize state growth/block).

These are two docs that contain really valuable feedback from members of the community regarding data collection they're interested in.
Would appreciate if we could mention this and get ideas from folks async.

This should help clear the doubts regarding EIP-7938 - State Growth handling.