Let’s talk about ASIC mining


#635

I don’t think any community member can just vote, you have to be on a panel first.

From the link you posted:

While the Community Governance Panel—those casting votes—will be limited in number, any member of the public can open a Pull Request for a new ballot item https://github.com/ZcashFoundation/Elections/blob/master/README.md

A foundation blog post says:

Around 200 members of the community will be selected to form what we’re calling a “Community Governance Panel.” Each of these members are in the process of being vetted by me and other members of the Foundation to broadly represent the Zcash ecosystem.
https://z.cash.foundation/blog/zcon0-and-community-governance/


#636

Ok, my mistake, 20cha


#637

Lets try this and I dont know if it will mean anything, but im going to explain (as non biased as I can) your idea of asic resistance to you and then you explain the daisy-chain idea to me and then we’ll check to see if we missed anything
At the very least I can reassess my position on it, I just feel as though I only have more metaphors for which to elaborate with
In the next few days, i was planning on another summary anyway


#638

Hey folks! I don’t even have time to write complete replies in here, much less read all of these hundreds of messages. :-{ I hope that someone is somehow boiling down the major sentiments and ideas for me and others to use. Maybe the Zcash Foundation’s voting protocol will serve for that purpose. Not sure.

Just jumping in here to say two things:

  1. I’m really chagrined that I let it sound like we were committing to a social contract of ongoing ASIC-resistance. That is absolutely never what I had intended to commit to, because (a) I always thought that it would probably become impossible long-term, and (b) I always believed that there was a fundamental trade-off between widespread distribution of the coins on one hand, and miners having a large sunk-cost investment into the coin on the other hand, and that the latter might eventually prove to be valuable for attack-resistance and network stability. (I still believe both of these things: even if ASIC-resistance is possible this year and next year, it will almost certainly become impossible given enough scale and enough years. And: distributing coins widely with a commodity PoW is fundamentally incompatible with locking miners into a sunk-cost-investment incentive-alignment with custom PoW, and the latter has advantages that might eventually prove to be important.)

    But, I allowed ZcashCo’s public statements to let people jump to that conclusion (i.e. we stated that our motivation was widespread distribution of the coins and the mining, which allowed people to reasonably assume that this was a long-term principle rather than a temporary strategy), and I have to take responsibility for that and if possible to try to honor what people thought.

  2. I’ve been trying to educate myself about the state of cryptocurrency mining. I’m getting in touch with various big mining operations to try to understand their economics. I’m also learning about energy production (direct contracts with plants instead of sucking from the grid, renewables, etc.). One major unanswered question in my mind is: suppose you have a large, profitable GPU-mining farm. Why don’t you reinvest all your profits in scaling it up? In other words, why haven’t professional, large-scale GPU-mining operations already driven small-time miners out of business?

    One thing I’ve learned along the way is that GPU mining is absolutely essential to Zcashers in Venezuela. If Venezuelans try to import ASIC miners (i.e. for Bitcoin, currently), then they risk having the miners stolen or extorted by the army which controls all imports. GPUs are not (yet?) nabbed on import like that, and anyway there are a lot more GPUs already floating around inside Venezuela. This really goes to the point that Simon has made a few times in this thread, that custom mining hardware is more vulnerable to special treatment by authorities.

Oh, bottom-line, by the way, is that I’m basically still in the same place now that I was four years ago when we first decided to go for widespread-distribution-of-coins at the expense of sunk-cost-incentive-alignment. I still think that widespread-distribution-of-coins is more important (but I still think that it can’t last forever, and that the other side of the trade-off may also prove to be important).


#639

Its been pretty fudless lately, i think
Edit- and these metaphors guys, like let’s dig another pool at the waterpark so when the field trip shows up we can all still swim
Without leaving the waterpark, without putting that silly floaty rope down the middle because you know they’re going to swim over anyways
I cant just turn it off! :stuck_out_tongue_winking_eye:


#640

That should be the most important reason to have ASIC resistance. ASICS more easily become centralized and more easily controlled/confiscated. When you are creating the most private coin out there it will become a target for authorities and you need a decentralized mining community that can use hardware that is not so easily controlled. I am not convinced ASIC’s make a network more safe in the first place. Is it easier to gather ASIC’s to perform an attack or GPU’s? I would argue the former since they are only good for one thing. As long as the coin is profitable to mine there will be miners to mine it. Any theoretical advantage gained by more hashpower really doesn’t make much difference since you have to have a majority to carry out an attack. Wouldn’t it be better to look for an alternative to attack resistance by using ASIC’s? What about Komodo’s approach to backing up the blockchain for Zcash on the Bitcoin blockchain? Then attacks have to be aimed at Zcash and Bitcoin. Maybe it is possible to back it up on more than one blockchain? In that way you have attack resistance plus distributed mining. I think that approach would further both of your goals and you wouldn’t have to choose.


#641

No its not unanswered, its business 101:

To expand a large miner must find more floor space, more power, expend more capital and increase their risk. This pushes out their ROI. With a large GPU farm its a race to profits, expansion puts you back into the red with a building full of 1+ year old GPU’s.

This IS exactly what many small miners do. They do not have any of the facilities costs that a large farm has. However, this is a long slow process for a small miner, and the ROI on a GPU can fluctuate significantly based on price and difficulty.

You have some misconceptions about mining economics. How does a large mining farm drive a small miner out of business? They both have access to the same GPU’s. The cost for power is more than offset by the cost for floor space, buildings, insurance etc etc. The business models are completely different and the profits margins are relatively the same, and can even better for the small miner.

Lets compare a large GPU mining operation with a modest small home based operation (say 10kW):

The large farm generally has lower power costs, but high cost for facilities. They tend to focus on the fastest way to get to ROI, so cheaper GPU’s, usually less efficient. If we look as the Nvidia P104, P106, and 1080ti, ~full out max mining, these numbers fluctuate a lot from day to day:

P106: daily mining $1.10, efficiency 2.39 Sol/Watt, ROI @ $0.03/kWh 194 days
P104: daily mining $1.66, efficiency 3.46 Sol/Watt ROI @ $0.03/kWh 237 days
1080ti: daily mining $2.75, efficiency 3.33 Sol/Watt ROI @ $0.03/kWh 249 days
Bitmain S9: daily mining $4.95, efficiency NA, ROI @ $0.03/kWh 384 days

So the primary risk for the big miner is getting to ROI as fast as possible. If the price crashes or the difficulty increases, the time to ROI skyrockets, so the faster to profits the lower the risk. I know of several large 2000+ GPU’s operations coming on-line and they are all using the P106. I don’t know any large farms that would even consider a more efficient long ROI GPU like the 1080ti.

By the numbers, a 352kW @ $0.03/kWh facility can support:
1,994 P106’s, startup cost of $637K, 1st year profit $82K PM 9.1%, 2nd year $537K PM 60.2%
1,994 P104’s, startup cost $995K, 1st year profit $126K, PM 9.7%, 2nd year $1.121M PM 86.6%
1,192 1080ti’s, startup cost $953K, 1st year profit $164K, PM 12.7%, 2nd year $1.117M PM 86.5%
213 Bitmain S9’s, startup cost $405K, 1st year profit $28K, PM 4.61%, 2nd year $607K PM 71.4%

So the 1080ti would make the best choice right? Not exactly, the 1080ti requires 47% more upfront capital investment, and the P106 reaches ROI a full 55 days before the 1080ti. For a 2000 GPU farm that is $124,300 in revenue before the 1080ti’s make a dime. That reduces your risk from increasing difficulty and fluctuating prices. This is also excluding discounts that large farms can get for volume GPU purchases on the P104 and P106 but not on a commercial GPU like the 1080ti. This will skew the numbers even more to the P106.

So now the small guy:

The small miner has higher power costs but no cost for floor space. So they tend to focus more on efficiency as they have a set amount of space and power available. A P104 is a mining GPU and has no resale value and limited warranty. The 1080ti on the other hand will still retain a considerable amount of resale value after reaching its ROI and still be under warranty.

a 10kW @ $0.09/kWh basement farm can support:
57 P106’s, startup cost of $18.1K, 1st year profit $955, PM 3.8%, 2nd year $18.1K PM 71.4%
57 P104’s, startup cost $28.3K, 1st year profit $2,213, PM 6%, 2nd year $30.5K PM 82.9%
34 1080ti’s, startup cost $27K, 1st year profit $3,301, PM 9%, 2nd year $30.4K PM 82.8%
6 Bitmain S9’s, startup cost $11.5K, 1st year profit -$569 PM -3.3%, 2nd year $11K PM 63.4%%

So the 1080ti has the best first year return and retains a significant resale value after reaching ROI. However, you need to understand that many small home miners don’t just go out an buy $27K in GPU’s, They may start with a $5K or $10K investment and then buy additional hardware with profits over time. So they typically have GPU’s well past their ROI and into pure profits as they buy new, this boosts their margins.

Then you need to consider the fact that small miners can push their mining efficiency way up if they choose. They can sell their GPU’s, and buy the latest more efficient models. Then you need to consider that 10kW is essentially free heat for a average size home in the winter months (I always move about 10kW of rigs in my home in the winter). All these things add up to keep the small miners profit margins even higher than I show here. So a large GPU farm cannot put a small miner out of business without first putting themselves out of business, even with triple the power costs. Small miners can weather high difficulty and low prices better than large farms as they have no fixed facility costs and can just turn up their mining efficiency.

ONLY ASIC’s can put small GPU miners out of business. I included the S9 so you could see how ASIC’s have impacted the economics of Bitcoin mining compared to GPU mining.

UPDATE:
Its important to point out that there are NO large GPU farms mining ZEC. None of the 2000+ GPU farms coming on-line have any intention of mining ZEC. You cant compete with ETH, and for large GPU farms its not about supporting the community, or the network, its about profits. Even for the 1080ti Bitcoin Gold routinely beats ZEC in profitability. The time to ROI and profitability dominate large GPU farm mining decisions. The brief few times that ZEC was more profitable than ETH you had some large farms pointed at ZEC and only then.

I find it ironic you don’t understand this. ZEC is currently 100% supported by small miners as large miners will not touch ZEC unless it is more profitable than ETH. These are the very people you seem so focused on driving away from ZEC.

Your lack of business knowledge about the mining economics of your own coin is simply frightening.


#642

You know the quality of forward-thinking is a much better adjective for the premier cryptocurrency then bound to rhetoric regardless of any kind of reassessment
I’m about positive I can find a historical situation where being tied to old thinking, regardless of its validity, did not end well
Edit- to say what we do now is almost entirely dependent on what was supposedly decided on what we thought to do then,
is dangerous
Edit- i guess that’s not an adjective but a quality


#643

https://shop.bitmain.com/product/detail?pid=00020180503154806494uGcSyiu806FD&utm_source=twitter&utm_medium=social&utm_campaign=z9-mini-announcement no longer a rumor and kicking 10K sols on 300w??


#646

They were mining ZCash as well. they just could hide well compared to monero.
As a supporter to zcash for over 6 months currently, I request zcash team to indicate clearly if they are still on their words of being Anti-ASIC and if so we need to know the counter-measures that they will take.


#647

Assuming Zcash team say Yes to ASIC… good for some people, bad for others.
Assuming Zcash team say No to ASIC. good for some people, bad for others.

Assuming Zcash team say no words. Bad for everyone.
So please, speak up.


#648

He needs to cut the shit and either step down, or give us a clear cut damn answer 1 way or the other.


#649

I am not disagreeing, but having an answer provides neede leadership on the topic. Which is what we would expect given the seeming departure from prior stated stance on this topic.

What we are getting instead is worse then my cat wanting to go out, then back inside, then back outside. It is frustrating, unprofessional, and lacking in leadership.

Step down or give us an answer.

If you say “nope, asic is okay with the zcash project and goals” then fine. I dont agree but at least you have the fortitude to make a decision @zooko.


#650

Why are you only concerned with them?


#651

Do you think we should Fork against these supposed equihash asics and if so why?


#652

They will when you see a rack full of ASIC’s in their plants. The technologies are currently at parity between the larger and smaller operations. That will be centralization. Can joe/jane miner buy quantity of 100 or greater ASIC from bitmain? No. Can a larger plant? Yes. Once they do, being a small miner will be a money losing proposition and small miners will exit. A few small miners will buy their allowed 1 ASIC, but that wont be even 10% of the current small time mining pool based on prior ASIC implementations (based on scrypt, cryptonight).

Either way. Make a decision.


#653

There is a possible scenario that scares me.
Assuming the Zcash team has a defined position Yes to ASIC and They don’t speak up THEN
This is privileged possession of information (like inside trading).

No one else without sure ASIC is ok (even large farms) will order the Z9 because if a fork is coming you lose all the money.

But people with privileged information will order it and will be devastating.

If this scenario is happening so this is really non honest. It is a fraud I say.


#654

So it’s here
https://shop.bitmain.com/product/detail?pid=00020180503154806494uGcSyiu806FD&utm_source=twitter&utm_medium=social&utm_campaign=z9-mini-announcement
Any word on the way forward?


#655

Wow, look @ the specs: Power consumption: 300W Hashrate: 10k Sol/s ± 5%

That would be equivalent to 25 GTX 1070’s costing over $9000, and power usage of 2500W.

No wonder it has been a pain in the ass to try to compete, they have been running these for months.

CHANGE THE ALGORITHM NOW until you figure a new plan . Let’s find out how much of the network hash Bitmain owns. Be different than Ethereum, please don’t just “do nothing and wait and see…”


#656

I’ve been mining Zcash the day it was released, i am at a cross-road.