Ethereum Core Devs Meeting 52 Agenda

[Souptacular]

Ethereum Core Devs Meeting 52 Agenda

Meeting Date/Time: Friday 4 January 2019 at 14:00 UTC

Meeting Duration 1.5 hours

YouTube Live Stream Link

Livepeer Stream Link

Constantinople Progress

Agenda

0. Quick announcement about note taking and Gitcoin bounty
1. Testing Updates
2. Client Updates
3. Research Updates
4. Working Group Updates
5. Constantinople HF
6. Istanbul HF / Roadmap
   • schedule for future updates
   • EEP-5
   • EIP-223 formal process of hard forks
7. ProgPoW

[5chdn]

From Gitter

Peter Pratscher @ppratscher 10:27
As the gangnam ProgPoW testnet is running smoothly since a few weeks & the mining ecosystem is maturing (open source implementations for cuda & opencl are more or less done, claymore (the dev of the most widely used closed source gpu ethash miner) has also confirmed to add ProgPoW support to his miner) I think it would be a good time to finalize the discussion if Ethereum should switch to ProgPoW during the next core dev call?

Thanks, ideally a go/no-go decision would be the best outcome, including a rougth timeline if it is a go

[5chdn]

Now, that Constantinople is finalized, I would propose a rough schedule for a subsequent protocol upgrade (a.k.a. "Istanbul"?):

• 2019-01-16 (Wed) projected date for mainnet-hardfork ("Constantinople")
• 2019-05-17 (Fri) hard deadline to accept proposals for "Istanbul"
• 2019-07-19 (Fri) soft deadline for major client implementations
• 2019-08-14 (Wed) projected date for testnet-hardfork (Ropsten, Görli, or ad-hoc testnet)
• 2019-10-16 (Wed) projected date for mainnet-hardfork ("Istanbul")

That breaks down to a fixed 9-months cycle to release protocol upgrades accepted prior to the hard deadline in May to mainnet. All proposals accepted after that date should go into a subsequent hardfork nine months later.

Action items for the call:

• Shall we go with a fixed release cycle in the future?
• Are nine months sufficient, or maybe too ambitious?
• Are we calling it "Istanbul"?
• ref EEP-5: Ethereum Hardfork Process karalabe/eee#5

[peterbitfly]

Regarding ProgPoW here is a summary of the current status (to my knowledge) of the development efforts:

• Parity and Geth have pending PRs that add support for the ProgPoW algorithm
• There are two distinct CUDA & OpenCL open source miner implementations: https://github.com/AndreaLanfranchi/ethminer and https://github.com/EthersocialNetwork/ethminer/
• On the 11.12.2018 a ProgPoW testnet titled Gangnam has been launched which successfully transitioned from Ethash to ProgPoW on block 3000
• The testnet stalled during the first ProgPoW epoch transition. The root cause for this issue was that the only active GPU miner on the network was using an outdated miner that crashed during the epoch transition. The issue was already fixed on a more recent version. Since then the testnet completed 3 additional epoch transitions without any problems.
• ProgPow support was added to Open Ethereum Pool (an open source ethereum mining pool) https://github.com/gangnamtestnet/open-ethereum-pool.
• Verification benchmarks: A well-optimized implementation should end up being about just 2x slower than the ethash verification as there is twice as much data read per iteration. (source Progpow integration go-ethereum#17731 (comment))

Please feel free to amend the list in case I forgot something important

[AndreaLanfranchi]

Just for sake of precision :

• https://github.com/hackmod/ethminer is quite behind
• @hackmod most recent implementation is https://github.com/EthersocialNetwork/ethminer/ which is also based on my work.
• My implementation is continuos WIP and going through optimization phase.

Worth to mention that with my work ethash and progpow live together nicely sharing the same dag data and smoothly switching from one algo to the other.

If ProgPoW is well prepared can't we just enable them on Constantinople Hardfork period with other EIPs? GPU miners wouldn't be able to earn profits with 2 ETH block reward without anti-asic implementation. ( See https://eips.ethereum.org/EIPS/eip-1588 )

ps) node is being ready on https://ethereumprogpow.com and we already have our miner implementation on https://github.com/gangnamtestnet/progminer/releases

[AndreaLanfranchi]

If ProgPoW is well prepared can't we just enable them on Constantinople Hardfork period with other EIPs?

Constantinopole hard-fork is scheduled for January 16.
There is no chance we can possibly do all the tests for ProgPoW in such a short amount of time.
Yes gangnam test network is working but :

• it's too small (at the time of writing only 10 nodes have announced)
• too low hashrate (31 Mh/s is one rig !!!)
• among major pools only Bitfly has actively participated

We still need to :

• test consensus on high number of nodes (the increased evaluation time of solutions have an impact)
• be sure all major pools and mining implementations do not face any problem on a large chain (1 million blocks at least)
• we have tested on very few epoch changes and periods
• test (with miners that can support algo switching) a simulation of fork
• and lots of other things ...

@AndreaLanfranchi I appreciate your comment, but I would also argue that, we already had a plenty of time and now is Jan. 2019, regarding this EIP it has been more than 7 months of development, QA, etc.

Since ProgPoW already have it's testnet I strongly suggest to enable ProgPoW with Constantinople.

[5chdn]

Doesn't matter, Constantinople is final. You are free to propose it for a subsequent hardfork.

@5chdn With EIP-1234? Lol it will cause more than 30% of nethash drop, I am very concerned of the future of ethereum....

[AndreaLanfranchi]

@naikmyeong yes that's the creation time ... but I remind you that it's never been officially accepted.

[AndreaLanfranchi]

Lol it will cause more than 30% of nethash drop

Recent price drop has caused more hashrate drop than that. And still ethereum PoW is more than secure. Difficulty will adjust as a consequence.

And since we are 2 weeks before constantinople I believe we can decide go/no-go of constantinople this time.

Recent price drop has caused more hashrate drop than that. And still ethereum PoW is more than secure. Difficulty will adjust as a consequence.

By banning all miners you mean, there will be no GPU miners left on the mainnet

[AndreaLanfranchi]

And since we are 2 weeks before constantinople I believe we can decide go/no-go of constantinople this time.

Give up any hope to see progpow shipped with constantinopole.
There is no time !