The ASICs are here. What does the community think?


#1

Bitmain just released the Z9 equihash miner.

https://shop.bitmain.com/product/detail?pid=00020180503154806494uGcSyiu806FD&utm_source=twitter&utm_medium=social&utm_campaign=z9-mini-announcement

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.
I want to know what the community thinks of this?


#2

I’ll just post here what I think are some reasonable pros and cons about the topic.

Pros of ASIC mining

  • Higher network security if increased hash rate is decentralized
  • More stable network hash rate will make for more consistent block times
  • Easier to optimize cooling, higher density mining, big and small operations
  • Speed of expansion if units are readily available
  • Minimal downtime for upgrades and repairs at scale

Cons of ASIC mining

  • Difficult to change POW algo if necessary
  • Increased risk of 51% attack from malicious companies/orgs with a monopoly on production
  • Catch 22, miners must upgrade their operations to include ASICs to prevent a 51% scenario
  • Possible malicious code built into hardware
  • Availability to low volume purchasers
  • Short Warranties
  • Long repair times
  • Resale Value

Pros GPU mining

  • Flexibility to changes in POW algo
  • More accessible to low volume purchasers
  • Many competing companies and distributors of various hardware
  • More options available to beginning miners
  • More distributed network hash rate
  • 1 Year or longer warranties
  • Warranty claims processed quickly
  • Resale Value

Cons of GPU mining

  • Lower potential network hash rate
  • Less stable network hash rate due to profit switching pools
  • Greater potential block time variance due to less stable network hash rate
  • More expensive cooling and power usage per hash rate
  • More downtime durring upgrades and maintenance at scale
  • Possible malicious code built into software

#3

Higher hashrate does not have any effect on network security. What gives the security is the power (in electricity) that is spent to obtain that hashrate. In other words, what matters is the number of Watt that is put in the blockchain, not the H/s. So I don’t see how the first point is any argument in favor of Asics.


#4

Hi Cyril. I interpret this to mean the resilience of the network is increased if it’s more decentralized, and the ASICs will increase centralization. I’m also wondering about the impact of using container technology to host secure nodes for the same reason.


#5

@DarkStar : I agree with that, but how is that a pro for Asics (according to the list of @Sto1cNate) ?


#6

Another argument:

Algorithm changes or improvements will probably be unsupported by ASICs and hence unwanted by the mining community. This could force zencash in a locked-in situation and slow progress.

GPUs on the other hand are more flexible. They can handle algorithm modifications more easily and without the incompatibility risk.


#7

Good point. GPUs do provide you extra flexibility if there does need to be a change in the algorithm.


#8

I will crosspost this from the github issue where I posted:
For large decisions that can have significant effects, I like to do a worst case/best case exercise. Identify the worst outcome of a decision, and balance it against the best case of the decision. It helps to actually quantify things in doing this as well.

Furthermore, part of this decision should be reviewing what the goals of the ZenCash project are and how the decisions to be made are consistent with the goals.

There are people making hardware purchasing decisions about this, and if too much of the entrenched hashrate comes from miners running ZenCash ASIC’s then over time there will be a group of people wanting to maintain that hashrate. Making a decision on the direction to go is something that should be made withing a reasonable time period.

Choices under discussion so far that I have seen:

  1. Do not change ZenCash Equihash algorithm
  2. Modify the ZenCash Equihash algorithm to require more memory by changing parameters
  3. Change the ZenCash algorithm to a different one already in use
  4. Change the ZenCash algorithm to a completely new one

One of the things I have noticed as a cryptocurrency miner is that the cryptocurrency that is the first to have a new algorithm in many cases becomes the lead of that mining class. It may be because the ones that have their own algorithms also have developers that continue to create and improve the cryptocurrency, so it may be more due to the long term effect of active developers, and the same effect could be obtained by continued active development on the project.

Some of the best discussion I have seen on the zcash github is in this issue https://github.com/zcash/zcash/issues/1211


#9

All Equihash-PoW coin community leaders are being asked this same question.

I think the underlying hope is that they will provide input and thought leadership so that mining operations built primarily on GPUs can respond correctly.

Sure, Bitcoin hasn’t disappeared with the dominance of ASICs- but ZenCash is not Bitcoin.

And SiaCoin isn’t totally disappearing with ASIC dominance either- but GPU mining of SiaCoin is all but abandoned, as even dual mining alongside ETH does not make financial sense.

So, perhaps a better question is, “Should GPU miners wait and see, or should they cut bait and turn their efforts to coins/communities that will protect their interests?”

Sadly, I think I just answered my own question.

But one answer is still necessary: what does the ZenCash team intend to do?


#10

I really hope ZenCash will fork off the ASIC equihash algorithm and will stay only GPU mineable. Because

  1. GPU mining makes it mining more decentralized (also lower entry costs to start mining Zencash)
  2. GPU mining makes Zencash more flexible with algorithm changes compared to ASIC mining, e. g. maybe changing to a PoS system in five years is an option, but ASIC miners would never approve to that change.
  3. 51% attacks are much more likely with ASICs, especially for for the coins with a lower network hashrate, which Zencash has compared to e. g. Zcash.
  4. Because of idealistic points. Bitmain would have a monopoly on making Zencash miners and they always mine with the new miners themselves and no third company should have such a power over Zencash which wants to be as decentralized as possible in every other area.

#11

Personally I think GPU mining is a more open and accessible method for community members to contribute to the hash rate. We have many members because it’s possible to make incremental investments in GPU power, and running these rigs is “home friendly”. They’re not horribly noisy, and resale value of GPUs is reasonable.

I believe that the Zencash network WILL survive even if we allow the Z9 invasion to happen, but the mining constituency will definitely change. Clearly the community will feel some pain because their investment will be devalued. Go and count the number of posts on this forum that request help with mining on GPU rigs. We have a thriving group of GPU miners here.

This is definitely a topic that deserves a community vote. There’s nothing in the Zencash whitepaper that says the project strives to resist ASIC hardware, however because it uses Equihash, a supposedly ASIC resistant algorithm, the community has a lot of supporters that bought into the belief that the network should be supported by a decentralized set of miners.

The Zen philosophy page states “Everyone has an equal opportunity to participate in decision making, which empowers stakeholders and makes ZenCash less susceptible to third party influence.” Clearly, Bitmain is a third party that could influence the project and we should voice our opinions. To be true to its charter, Zencash should AT A MINIMUM allow a vote on the matter.

But, we don’t yet have a system that lets us vote on this issue. How inconvenient! Here’s a quick proposal that we can follow to send a signal to the leadership:

  1. Two weeks from this post, I propose that Secure nodes that support ASIC resistance shut down for 24 hours. Do not take your rewards for Sunday, May 20th if you want the team to consider options to change the POW algorithm. Leadership should count how many nodes they had active on average the week before Sunday, May 20th, and how many nodes were rewarded for that day. I propose the two week delay to allow Rolf and Rob to communicate this ad-hoc vote on the bi-weekly call and get the message out.

  2. On that same day, to capture miner sentiment, do not mine any Zen. Switch to the coin you would mine instead. Lets check the network hash-rate and see how much it drops. Yeah, I know people will take the opportunity to cash in on the low difficulty potentially… so be it. Afraid of a 51% attack? Well, this is how many miners there will be if the portion of the community that believes in decentralized hardware stops mining. At least the ASICs aren’t there to skew the numbers yet (we assume). Do you want your GPU rig to be useful for mining Zen after June 2018? Shut it down for a day and save your investment. Otherwise, be spend the day at least researching an alternate coin that you’ll be forced to mine.

  3. I’d also like to propose that users abstain from using Zen on that day, but I am not convinced that the measurement of the transaction count would be reliable, since many transactions will occur regardless of user activity, paying nodes and miners. I’d much rather see a VoteCoin proposal to capture user sentiment. I’d like to see the leadership organize that.

Assuming that ASIC resistance is important, there are going to be plenty of options that can be voted on later.

I suggest that we save energy and follow Bitcoin Gold’s (TBD) approach. Bitcoin Gold is dedicated to staying ASIC resistant and are using Equihash. If they are very forward thinking, they will install a regular update policy similar to Monero that modifies POW on a regular basis. At a minimum they should come up with a stop-gap solution before June. If they don’t, it’s pretty much curtains for Bitcoin Gold. It seems logical that a consortium of ASIC resistant projects should arise from this, but we’ll see how willing the crypto community is to join forces.

Of course other possibilities are available, and there is no reason to rush into anything, but that’s not the focus of this post. My number one message here is - let us voice our opinion on the matter. If the leadership has a voting proposal, lets do that. Otherwise, the community can and should speak to them in alternate ways. Secure nodes and miners, consider the May 20th proposal.


#12

Here is a link to the current discussion on the zencash github https://github.com/ZencashOfficial/zen/issues/142


#13

There is only one acceptable solution here, and it’s not acceptance of ASIC miners. This is also not a decision you should be taking your time with. Yes, we have a month and half before Bitmain ships anything but it’s now when people are making purchase decisions. Make the call, make it public and move on. We all know there is no positive with allowing ASICs to mine any currency.


#14

I’d like to note that profit switching pools also exist for ASICs.


#15

Here is a link to ongoing work on basic resistance for monero


I wonder if the zencash devs can look into this approach?


#16

Wow where to start. First of all, I am strongly against ASICs mining. Full disclosure, i do own few rigs, total hashpower 5800 Ksol.

First, I would like to start with Z9 mini. A lot of people assume, as well as I do, that Bitmain already used these miners since January. Someone mentioned even seeing dust on the fans, they didnt even bother to clean it for a picture. Nowhere on the website it says its brand new. As it is a thing with used products you get limited warranty - 3 months warranty since day of shipment. Shipments from China usually take long time at least a month - month and a half. Which means miners are left with month and a half warranty on second hand merchandise compares to warranty you get for GPU’s which is usually 2-3 years depending on the brand.

Secondly the name Z9 mini - which means there is already a product in their palette that is not mini. Z9 if you will, which uses 1000-1500W power like its usual for ASIC miners. Guess the hashpower on that one? That is right, it will make your Z9 mini obsolete. My guess is somewhere around 30Ksol-45Ksol. These ASICs will not be on sale till Bitmain makes even better ones, and then he will sell you his used ones.

Third point in my post is this. If Bitmain has in fact had these ASICs since January and has even more powerful ones now, he can stock up large quantities of ZEN, be that for dumping or for staking for secure nodes and supernodes. Either way it gives him too much power over the network for voting over issues and 51% attack.

For this reasons we need to fork and fork as soon as possible like Monero did. It is clear that Zcash will not fork by reading Zooko’s reply’s on Zcash forum. Zen can LEAD, FOLLOW or GET OUT OF THE WAY.

In case of fork i see the following scenario happening:

Zcash doesnt fork, while Zen does. Most of the GPU power will move from Zcash to ZenCash. ZenCash will become community leader along with Monero (which has AMD rigs pointed at it, while Zen would had mostly Nvidia). Zencash difficulty would rise, but so would it value as its harder to mine it, while still being decentralized. Secure nodes and SuperNodes would become even better way of earning ZenCash as its harder to mine it. If we compare ZenCash to Price Zcash had last year when it was GPU community leader along with Ethereum for mining, it had 40$ value in April and went up to 878$ in its ATH price. If ZenCash should fork and Zcash should not, we can expect at least 800$ by the end of the year for ZenCash. Imagine gains on SuperNodes and SecureNodes on that value.

I understand that some of you are impatient in waiting for team to announce their final say, and have bought Z9 mini, and you will probably now vote in behalf of ASICS just to cover your investment, but you guys will still be able to mine Zcash.

For those of you who havent bought it but are thinking, please take this in consideration. You order a 2000$ ASIC and with current difficulty you can mine aprox 40$ a day. This will change if we let ASICS run the world, it will go down to 5-10$ a day as soon as people get their hands on them. This is a ROI of 200 days, without power cost, to which I suggest buying the coins you wanted to mine instead of filling Bitmain’s wallet, and barely getting your money back. Sure the value of coins could go up, but it could also go down because of lack of trust from the community backing it so far.

For those who think ASICS are a way forward, they are not. They are a step backwards towards Centralization.
rotch

Well If I think of anything else I will post it, but please take these things into consideration.


#17

so everyone can judge for themself here is their offical twitter anouncement https://twitter.com/BITMAINtech/status/992034662875779072/ i can clearly see dust on that fan


#18

The risk of malicious code in the Z9 is incomparably higher than the same for GPUs. For one, no one has caught GPU drivers calling home. GPU producers have no benefit at all from risking their main (gamers) market by introducing dodgy code, but Bitmain can hold coins that have adopted its hardware hostage with an ability to turn off the hash power at will. And with Bitmain it is not a discussion of whether they would add such code as it is already demonstrated the Antminer is backdoored.

All CONS are not born equal either… a difficulty to change PoW algo is an inconvenience. The risk that a single company out there can blackmail the coin with an ability to switch off the vast majority of miners is a show-stopper. Yes they will hurt their business if they do it, but once a backdoor is created, there is no guarantee who will get access to it. Right now, the guys at antbleed are saying the exploit in the antminer is unauthenticated, so anyone who can take control of the home domain or intercept the requests (through a MiTM attack) can use the backdoor to shut down antminers.

Because of all that I think you should remove the malicious code from the GPU cons list, and highlight it in red in the ASIC list. Everyone commenting on this thread should visit antbleed.com and (excuse my bluntness) pay attention while there.

Finally, no one should take this post as a call for immediate fork - panic is rarely helpful. But we do need to find a way to stay ASIC resistant.


#19

Order limit per user raised to 50.


#20

I will not remove the con from GPU mining. Malicious code in place of mining software is a serious concern. In no way did I suggest the GPU manufacturers are hiding code in the driver software or other first party software. Closed source and even invalidated open source mining software does put yourself at risk.