Back to all reports

以太坊研究日报

Ethereum Research Daily

EIP-8141 received a PR raising MAX_VERIFY_GAS from 100k to 300k, based on Nethermind's DoS benchmark and measured gas for privacy pools and post-quantum signatures. Another PR clarifies that blob fees can lower the GASPRICE of a frame transaction. Marius and Ethlabs gave Glamsterdam updates: the goal is to have audit reports before ACDT on 11-16, and neither spec changes nor Hoodi have been decided. Versioned Opcodes became EIP-8488 and got a discussion thread. ETHBEAT launched its CROPS guide.

Image for November 10, 2026

Image: GitHub, ethereum/EIPs PR #12644 Update EIP-8141: raise MAX_VERIFY_GAS to 300_000

EIP-8141: validation budget and fee semantics

EIPs PR #12644: Update EIP-8141: raise MAX_VERIFY_GAS to 300_000

Opened at 08:58. The reasoning: 100k leaves out validation that real apps need. A Groth16 privacy pool spends 236,285 gas in its VERIFY frame (216,463 of it on the proof check), with Nethermind and ethrex agreeing to the gas unit; post-quantum signatures cost 94k to 276k. The benchmark (ethresear.ch 26224) measured caps from 100k to 500k: at equal failing gas per second the harm to a node is the same at every cap, and the cap only scales how much one peer can push. At about 100 failing transactions per second per peer, 300k keeps the added block-building delay within 25ms on local builder hardware. Nethermind already defaults to 300k for its own mempool. The PR is independent of #12301 (whether the value is a floor or a maximum).

https://github.com/ethereum/EIPs/pull/12644

ethresear.ch 26224

EIPs PR #12644: Update EIP-8141: raise MAX_VERIFY_GAS to 300_000, opened at 08:58. The reasoning: 100k leaves out validation that real apps need. A Groth16 privacy pool spends 236,285 gas in its VERIFY frame (216,463 of it on the proof check), with Nethermind and ethrex agreeing to the gas unit; post-quantum signatures cost 94k to 276k. The benchmark (ethresear.ch 26224) measured caps from 100k to 500k: at equal failing gas per second the harm to a node is the same at every cap, and the cap only scales how much one peer can push. At about 100 failing transactions per second per peer, 300k keeps the added block-building delay within 25ms on local builder hardware. Nethermind already defaults to 300k for its own mempool. The PR is independent of #12301 (whether the value is a floor or a maximum).

https://ethresear.ch/t/how-large-can-eip-8141s-validation-budget-be-a-mempool-dos-benchmark/26224

EIPs PR #12642: Update EIP-8141: note that the blob fee can lower GASPRICE

Opened at 07:45. The fee budget section says GASPRICE excludes blob fees, which reads as if blob fees never affect it. In fact, when `max_fee` does not cover the full tip, `max_priority_fee` is what is left after `required_max_fee`, and `required_max_fee` includes `blob_gas * blob_base_fee`, so a higher blob base fee lowers GASPRICE for the same transaction. This was found while checking Nethermind's frame tx `effectiveGasPrice` against the formula.

https://github.com/ethereum/EIPs/pull/12642

Glamsterdam: this week's timeline and what it means for accounts

written by Barnabé

Ethlabs: Ethereum gets stronger when its people ask hard questions, 16:13, written by Barnabé: Glamsterdam can raise Ethereum's capacity by more than 3x, but developers have raised important concerns about some of the proposed changes, and Ethlabs spent the past week working on them with the EF, auditors and core developers.

https://x.com/barnabemonnot/status/2120084651343753250

Other updates

Community