Ethereum Core Devs Meeting 46 Agenda

[lrettig]

Ethereum Core Devs Meeting 46 Agenda

Meeting Date/Time: Friday 14 Sept 2018 at 14:00 UTC

Meeting Duration 1.5 hours

YouTube Live Stream Link

Livepeer Stream Link

Constantinople Progress

Agenda

1. Testing
2. Client Updates
3. Research Updates
4. Constantinople a. Client progress. b. Ropsten block number. c. Can we do it before Devcon?

Time-allowing:

5. EIP-1380: Reduced gas cost for call to self
6. EIP-1108: Reduce alt_bn128 precompile gas costs
7. EIP-1057 (ProgPoW) and EIP-1355 (Ethash1a)

[mdaria510]

      Seriously. A lot of people are NOT happy with the broken promises. If you want to push your most staunch supporters away at a time where every centralized shitshow is claiming to be the next ETH killer - go ahead. Weird, I feel completely helpless as both a miner and a user so who really has say? Because it’s not the community. Maybe the initial ICO investors?

Edit: I watched Zcash hashrate fluctuate from 1.05gh to 1.61gh today. Which means one entity (likely Bitmain) could easily pull off a double spend off any major exchange. I know what you are thinking - that’s not half. Well they don’t need half if the other pools are split up. They just need the longest PoW - how much high than 50% dictates how far back in time you can reorganize. When these new ASICS come online there will be a time where no miner can “defend” against this security break since we won’t have to proper power to Claim longest POW. The exchanges stopped trading one time - will they do it again when several hundred million are stolen? Or will they just delist us?

You shouldn't feel helpless as a miner. You are empowered as a miner. Don't underestimate how much hash power GPU miners can collectively wield if someone were to effectively organize even a small percentage of them. We don't necessarily have to have the longest chain. We just have to have enough power to credibly back the existing chain and force exchanges to consider listing both chains, and insist that our chain is called "Ethereum", and nothing else. There needs to be enough of an uproar that press starts spreading the word that Ethereum is struggling to keep itself together while the price is already tanking. That alone will be enough to further tank the price to counteract any supposed benefit from the issuance reduction. It needs to be a no-win situation for the devs until they compromise. Think back to segwit2x - up until the cancellation there was enough theoretical support on both sides to force exchanges to consider the possibility that both would exist. No one ever had to mine a fork, they just had to believe that it plausibly could happen.

@MoneroCrusher As soon as you fork and rename it you've lost. It's not an option. The only plausible way forward for GPU miners is to "go on strike" and commit to continuing to mine the existing chain and undermine the unity of Ethereum and force compromise.

Because it's pretty clear that the devs consider us as second class citizens and not their most devoted users. It's time to remind them what decentralization actually looks like.

Devs don't control the blockchain. They issue proposals in the form of code, and users and particularly miners choose whether or not to approve those proposals by running that updated code. I do not approve of this proposal, and I will not run this code.

If there are pool operators out there that will publicly commit to continue running the existing chain past the fork, please step forward now and I will do everything in my power to spread the word. You have a good incentive to do so - it would make your pool unique and attract the attention of GPU miners worldwide, and you would therefore benefit from a significant increase in income prior to the fork, no matter what happens. I would do it myself if I wasn't prohibited in my state.

In either case if there's no movement on this before the end of September, I'll personally make it a quest to ensure that every and anyone with an incentive to take Ethereum down a peg or two is aware of the situation, and I'm sure it won't be terribly difficult to find at least a few wealthy individuals that will throw some resources at it, whether its hash, mining pools, propaganda, or whatever. I can already think of more than a few that would love nothing more than to see Ethereum crash and burn.

[Nmining]

From media https://cryptonewmedia.press/2018/09/18/how-much-did-asic-impact-crypto-mining-a-look-into-constantinople/

[MoneroCrusher]

@mdaria510 That's a far greater idea than mine. Only when shit hits the fan we could think about it. In your "resistance" scenario all that miners have to do is nothing. This is great. We need to get some pools on our side because the pools will also get less revenue since ASICs & big farms will mine on their own private pools. I for sure will not provide any hashing power to the new chain.

Edit: That is if no plan to address this issue is laid out by the EF this month. And no, Istanbul is not an option to me and many others. This needs to be implemented shortly after Constantinople.

[5chdn]

This needs to be implemented shortly after Constantinople.

Then start working on it. I'm happy to accept any ASIC-resistant engine and tests for Parity Ethereum.

[MoneroCrusher]

@5chdn Would me getting a private testnet running help you with that? I got instructions on how to set one up with ProgPOW. I'm not a programmer though, but as far as I know OhGodAGirl is already in contact with you guys.

[theWUJustin]

I would agree, @eosclassicteam. It's not in Constantinople, but there's nothing preventing it from being implemented soon after -- if there's dev support, and it's considered a high enough threat/priority.

I think this is a very high threat against Ethereum, for the following reasons:

1. GPU miners are being pushed out of the market. This leads to higher levels of centralization with ASICs. Of course this is bad for one of the central pillars of blockchain, which is robust decentralization.

2. We have many instances of ASICs completely dominating a coin and forcing them to make policy decisions at gun-point, and I don't think Ethereum core devs want to put themselves in this situation.

3. The argument "ASICs aren't a threat because POS is 'around the corner' is a failed argument considering that by the time even the first implementation of POS rolls out, the E9 miner will already have shipped and be mining, and most GPUs will have dropped off the network. This makes transitioning to POS much harder since you'll be fighting all the ASICs with their break even timelines who won't want to switch. That will cause another divide in the community and result in an "Ethereum Classic Classic The Remix Part II The Sequel feat. The Zombie Doppelganger Twins" which is a disaster squared.

ProgPow and EIP 1355 should be a priority because:

1. You literally have people like OhGodAGirl offering to do all the work, and the attitude is basically "we don't need your help". That seems very arrogant and is not in character with the open source community where everyone tries to help each other.

2. The GPU miners are pissed that they are getting a block reduction down to 2, and not changing the POW to address ASICs is like a slap in the face. We need to get them back on-board, or else they will mine other coins just to spite ETH, and we are seeing that already with threats of this, and also spreading FUD, which brings me to my next point.

3. Endless sources of FUD surrounding this will happen and further tank the price. Ethereum cannot be successful at any stage if the ETH price is low. Investors will flee and interest will wane, and if there's only the hardcore developers there building "for the tech, man" you'll never get the holy grail of blockchain that everyone is supposedly building for: Mass Adoption

Thanks for reading

[holiman]

You literally have people like OhGodAGirl offering to do all the work, and the attitude is basically "we don't need your help". That seems very arrogant and is not in character with the open source community where everyone tries to help each other.

@OhGodAGirl has been very helpful, and nobody (afaik) has said "we don't need your help". But it's a fact that switching consensus engine require quite a lot of work:

The TODO list for ProgPoW adoption: https://ethereum-magicians.org/t/what-has-to-be-done-to-get-progpow-on-ethereum/1361.

[HiddeHoogland]

Why aren't we pursuing the decentralization that we intended Ethereum to be?

Having 6000 running GPU's on the Ethereum network for a while, i never thought we would need this discussion. My hashrate will be supporting XMR for the time being.

[mdaria510]

@holiman

I’ve lost count of how many times the roadmap has slipped because of things taking longer than expected, so it’s not like devs are in some double bind where there’s an immovable deadline and they’re being forced to do the impossible. The deadlines are self imposed. So they need to push the fork back if they have to and get it done.

Ethereum mining is big business now. There are millions of GPUs mining ETH. Which means billions of dollars of investment is on the line for people. Some of that is hobby miners but a lot of that is people who have made very, very significant investments in running the software they write. What devs consider an inconvenience and extra work is threatening people’s livelihood. Devs actually have the choice to implement this or not - anyone’s who’s livelihood is at stake doesn’t even have a choice, the only logical course of action for them is to oppose and derail the fork. I don’t know if that’s getting through to them - for people heavily invested in ETH mining on GPUs, ASIC resistance isn’t a “nice to have”, it’s a necessity. They can’t just wave this away and expect people to acquiesce because they don’t consider it a high priority. It’s the only priority for many people.

I also don’t want to hear any excuses about the difficulty bomb either. That was a cute idea when ETH wasn’t far and away the primary source of income for GPU miners. That idea worked when there was something else for millions of GPUs to mine, because then any miners that didn’t agree would just shrug and mine something else. That doesn’t work in a bear market where most miners are barely scraping by. It’s a gun to everyone’s head now - but GPU miners are the only party here backed into a corner with no other option but brinksmanship.

[lrettig]

it’s not like devs are in some double bind where there’s an immovable deadline and they’re being forced to do the impossible. The deadlines are self imposed. So they need to push the fork back if they have to and get it done.

This is not entirely true. The difficulty bomb is the immovable deadline. Call it an "excuse" if you like, but it requires us to hard fork relatively often, relatively on schedule. It's doing its job quite well.

[mdaria510]

It requires us to hard fork if we want to keep the chain running. But GPU miners are now being put in a position where there is effectively no difference between Constantinople making their hardware and investment obsolete or the difficulty bomb making their hardware and investment obsolete. I'd love nothing more than to see it removed because the only reason it exists in the first place was to get miners to act against their own self interest. Vitalik has said as much himself. Maybe they didn't forsee the perfect storm of a rally that brought in millions of miners only to have the price dump shortly thereafter, ultra efficient ASICs that made a joke of ETH's ASIC resistance and being the only chain keeping those millions of GPUs afloat - but these are the consequences of the game theory inherent in that ploy. The road to hell is paved with good intentions.

We both know it's no big deal to remove it in a fork, so it is unequivocally a self-imposed deadline. It requires a fork to keep Ethereum functioning, but it says nothing about what that fork contains beyond than an extension or removal of the bomb.

[MoneroCrusher]

Good news: https://twitter.com/nicksdjohnson/status/1042382028837203968 Nick Johnson now supports ProgPOW after chatting with @OhGodAGirl and wants to implement it! Hope my reddit thread helped accelerate that happening!

Timing is critical too, though. Devs knew about this since May. I think to restore miner community faith it should be done as fast as possible, not in Istanbul HF. Why not in Istanbul HF? Because the network will punctually centralize with Constantinople HF at the end of October. Many GPUs will be turned off or turned to other coins and new hashrate growth will only come from ASICs.

So when Istanbul HF comes closer and they want to implement ProgPOW then, we will have even more of a community chaos, since some people will have switched to ASICs by then.

Edit: I think it's worth pushing Constantinople back 1 month to be able to implement it, if this is deemed urgent.

[5chdn]

Edit: I think it's worth pushing Constantinople back 1 month to be able to implement it, if this is deemed urgent.

My estimation is ~12 months. Instead of pushing back Constantinople for another year, let's just do another hardfork.

[MoneroCrusher]

@5chdn According to @OhGodAGirl 's last claim she said she could easily do it until Constantinople. What makes you think it's going to take 12 months? She has repeatedly said that the hard part is already done. Only thing that's left is basically client work and that she would provide manpower for it, free of charge.

[nootropicat]

@MoneroCrusher

Getting ASICs to the masses prevents Ethereum to fork because community doesn't want to when they own ASICs. This protects their investment.

If they manage to sell ASICs to the masses, ASIC mining is no longer centralized, making the decentralization argument moot. Asic mining behaves as a physical proof of stake, offering much higher security than gpu mining (for the same block rewards).

[MoneroCrusher]

@nootropicat yeah unless they sell the ASIC that is 1:1 the efficiency of GPU to the masses for cheap and keep the one that is 8x more efficient to themselves, which is what they do. Alternatively check out the Zcash problem or the Dash problem.

[holiman]

Well, if it's "ok" to push back Constantinople one month, how can it not be acceptible to roll it out in a HF one month after the non-pushed Constantinople?

I would also support a switch to ProgPOW, but I do not think we should cram it into Constantinople. The implementation of the algorithm is one aspect, but there are other things that need to be done to ensure a switch goes smoothly, and those things needs testing.

My opinion is that we should consider a dedicated pow-switching HF shortly after Constantinople.

[nootropicat]

@MoneroCrusher there are at least two ethash asics providers now (Canaan and Bitmain) and they're competing, making that scenario unlikely and temporary at best. Keeping best products from the market invites new competition. If you think it's a realistic long-term danger, why that's not the case in bitcoin?

Alternatively check out the Zcash problem or the Dash problem.

That's too vague. What 'zcash problem'?

[MoneroCrusher]

@nootropicat With Bitcoin it's becoming less and less of a problem but it still is one (because after many years now we are finally seeing some competition). But Constantinople is coming soon and there won't be any free market competition (competition for the Ethereum chain between manufacturers at most) until then. They'd milk ETH as much as they can then create some FUD in the end and sell their miners to gullible people. The problem with Zcash ASICs is that the manufacturer already has a better faster generation and is now smartly phasing out the gen1 ASIC to get as much out of customers as possible. They started out at $10k and 70-80 break even days in June, in the beginning of September they were $3-3.5k (same manufacturer), also with 70-80 days break-even. Guess where we're at? 300 days for the june batch (assuming it was mining from June until now) and 180 days break even for the newest one. Those things will never break-even. They're not designed to. They're only designed to deceive gullible customers and generate revenue & off-roll risk for the manufacturer who is already mining on gen2. As soon as the last gen 1 is sold and gen 3 is manufactured they'll do the whole thing over with gen 2. Zcash miners are fucked over for years. Their governance completely failed them.

@holiman I don't care, it can also be a separate hardfork. I feel like if it's included in Constantinople it'll get done in a faster manner. And more importantly because you make yourself susceptible to 51% attacks in that time, especially when going from 2 ETH / ethash chain to 2 ETH / ProgPOW chain. Think those ASICs will just sit idle and do nothing while their value goes to 0? So in everyone's interest besides ASIC manufacturer's it's better to do the thing in one swift switch and be done with it. Don't forget: there are hundred of millions of dollars at stake.

tl;dr doing a separate HF after issuance reduction will increase drama unnecessarily. Either do it with Constantinople OR take issuance reduction out of Constantinople and introduce it in the ProgPOW fork.

[mdaria510]

@holiman For various reasons that should be obvious enough that I don’t need to list them, I have very little faith in it actually being implemented in a timely manner (or at all) once the pressure is off.

[Nmining]

Grabs coffee cup from table. So what is going to happen to Etherium and stuff and is Ripple going to kick Etherium down to position 3 today? Oh!! #### this is cold...

[theWUJustin]

@holiman I also agree about putting the transition to ProgPow in Constantinople. By reducing issuance to 2 ETH per block, suddenly many GPU miners are not profitable and will switch off the network. We've already seen a hashrate drop. This is the perfect opportunity for ASICs to sweep in and control the network. By delaying Constantinople to include ProgPow you don't risk anything except maybe a FudDesk article "Ethereum misses key hard fork deadline." or some garbage click-bait like that. Not a big deal.

But not including it in Constantinople you risk ASICs owning the network, which opens the door to a very contentious HF "soon after" Constantinople if it happens at all before it is too late. Nobody wants Part II The Sequel of Ethereum Classic Classic.

[lrettig]

Closing in favor of #58