Let’s talk about ASIC mining

It’s true, from 5am to 7am it was 20M. thanks nec.

yeah, it was 20M, for some time, considering the price of zec is not particulary interesting.

so imagine how low will your profit go with 50M diff.

and this numbers coming soon.

network just being raped.

As there’s a lot of jumping from coin to coin. I tend to look at daily average & total network hashrate for zec and then total hashrate across all equihash coins for my indicators.

When you plot out hashrate across all equihash coins the graphs are less chaotic.

Mining is running computations. There is no disclaimer as to GPUs can’t be used for such a workload. GPUs are designed to perform such tasks. There is no restriction to not run a GPU 24x7. That’s the reason mfg (with good after sales service) don’t ask such questions. Hardware abuse with overvolt and heat were the primary reasons for GPU failures. Nowadays, the limiters on GPU do a good job to prevent it.

2 Likes

I could provide hundreds of additional FACTS on this subject, I did this for a living. In all honestly I don’t think that matters, or presenting any more FACTS would do any good.

I just want to say that @ZC93 does know more about this subject than myself and that, self-servingly, I now expect my Z9 mini to last for at least 10 years. :wink: I have a lot of respect for your experience in the industry, no doubts there.

Perhaps my logic is wrong, but this seems really simple.

A new ASIC has two possibilities: it could have been pre-mined or not pre-mined.

Given that an early failure rate is 1-3% or less given any new ASIC produced, I would still prefer the new ASIC over the pre-mined one. My reasoning is influenced by the fact that that I’m not buying hundreds of these so the chances of me having one that might fail early is less. If I were buying hundreds, maybe I would want them burned-in to lessen the chances of spending my time dealing with support, shipping and downtime.

From a manufacturers perspective, pre-mining may be smart, but in that case I’m wondering why Bitmain isn’t also doing it for support reasons rather than revenue reasons. Perhaps the test units they receive are a good enough sample size to determine failure rate potential, or maybe they are just taking their chances in order to get to market faster. Maybe we can expect Innosilicon’s pre-mined machines to last longer over the entire sample of units. For a single unit, the odds are so low of getting a failure that there are larger issues like potential damage in shipping or environmental variables that outweigh the mental energy spent on pre-mine vs non-premine.

The lifecycle decrease of a device (or silicon chip) due to pre-mining is irrelevant.

If they pre-mine and they sell it at small discount while stating on the website that it’s a used device then nobody would have an issue because then it becomes the consumers choice to purchase the used device. But if they sell used devices as new claiming that “Used is almost the same as new so we aren’t going to tell the customer!” is a deceptive businesses practice.

4 Likes

Buying fewer items doesn’t decrease your chance of receiving a broken, or soon to break item. If the failure rate is known and fixed, you are subject to that same rate. When you only have one or two you end up concentrating the risk of lost mining time as a guy buying 100 and losing 1 to infant mortality is 1% of his investment where it could be 50% of yours. Devastating difference.

1 Like

Buying fewer items doesn’t decrease your chance of receiving a broken, or soon to break item

If there is a 1% chance of getting a failed ASIC for each unit purchased, and you buy one unit, your odds of getting a failed ASIC are 1%. If you buy 100 units, the odds are closer to 100% that at lease one will fail early. The amount does matter because the probability is set by the total units at the time of purchase.

That’s not actually how the math works.

Please explain your reasoning?

Lets say you have a bowl of 99 plastic balls and 1 identical wooden one, totaling 100. If you select one ball out the initial chance that it will be wood is 1%. If you take the whole bowl, the chance is 100% that you will get a wooden ball.

If there is 1% of failed devices, the chance of getting at least 1 in 100 units is 1-(0.99^100) ~= 0.6339… which is equal to ~63.4%.

2 Likes

For your example you are correct. But that’s not an accurate model of the probability.
You’re actually taking one ball out of the bowl, determining if its plastic or wooden, and then placing the ball back into the bowl, shaking, and sampling again.

@MikeZhukov’s numbers are correct:

1-(0.99^100) for 100 units = 0.63 = ~63%
1-(0.99^1) for a single unit = 0.01 = ~1%

I think that’s true but only half the story of the formula and i would add this one:

If the one with 100 units purchased gets 1 unit he has a loss of 1%.
If by purchasing 1 you unlucky get the 1 unit out of the 100 you have a loss of 100%.

Means the guy with 100 units purchasing has about 0.000001% chance of getting a 100% loss, while you have a 1% chance of a total loss of investment. Not that it matters much in our risky business, but still …

Let’s just all stop talking about ASICs please, they are here and not going away anytime soon. PoS or something else opposing PoW may ultimately be the only solution. Until then, buy or mine aMaP. That’s why I have a dozen Z9mini, only plan for a maximum operational time of two years, and buy the dips. It’s sound enough to recoup the investment and have fun along the way. Best case what is accumulated now will be 10x more useful in a couple years or have enough to be meaningful for POS. Worst case, Zcash implodes, which is highly unlikely (so long as the manikins don’t take hold). It’s easy to get caught up in the profitability and hardware debates now, but it’s not conducive to the ultimate goal. One day we would like to securely trade Zcash for valuable products and services. If this can be achieved it’s a no brainer the value relative to the USD and all other fiat currencies will improve. It just so happens now you’re using fiat on a daily basis, so are more concerned with that relationship; one day that correlation may be extinct. Support the network and usage in whatever fashion is best for you and within your financial limits. If you believe in the future function of Zcash there is no discussion about current profitability, ASIC vs GPU, hard forks for ASIC resistance, none of that. If you would like on board to mine, then hash at it - if not, trade fiat for ZECs. The ASIC vs GPU debate, and should I shouldn’t I contemplation, is so over baked it’s painful to witness. We all know PoW will ultimately leads to a higher level of centralization than we would deem acceptable, so let’s discuss how best to implement a solution. Keep in mind though, it will take years to judiciously and beneficially implement such a change. Nonetheless, still time better spent than the persistent yammering.

1 Like

If you dont like to read just simple dont read…there are other topic on forum …great for you that you had money to buy lot off ASIC Z9 but it does not give you right to silence others discussion.

1 Like

Yes, it is great for me, thank you. Now look at the context of your reply in comparison to the post in question; it may become apparent how I achieved such greatness.

Please note, you have not been silenced. If my opinion leaves you speechless it was more effective than anticipated, but surely not the intention.

Whether used for developing grounds of what to do, or what not to do, everyone’s opinion is equally valuable. Yours included.

1 Like

@T.K

The problem with ASICs for Zcash is that Zcash claimed to be ASIC resistant and claimed that it would fork if ASICs were developed. Not forking means the team is dishonest, and if we can’t trust them with something as simple as tweaking the algorithm, how can we be confident that the trusted ceremony was honest? How can we trust that they don’t have backdoor to reveal private transactions?

So whether you think ASICs are good, bad, or neutral is beside the point. The Zcash team needs to fork if they want to retain an ounce of integrity.

2 Likes

Good point. But I don’t believe zcashco should be left to blame, or need to hastily fork to retain GPU miners. They do great work, and very pragmatically, which beside the applications of their efforts, could be the only reason I’m invested. It’s unfortunate an ASIC was developed to circumvent the equihash memory requirements, I was surprised, though not surprised enough to quit. It’s a natural evolution of hardware.

I doubt there is a ‘backdoor’ to reveal private transactions. Seems like a strong allegation doesn’t it?

Or it could be seen as a result of a poor choice of Equihash parameters, which failed to create the high memory requirements needed for fair mining between GPUs and ASICS.

The Zcash CTO even said they would change the Equihash parameters in a blog post from 2016:

“We already intend to alter our mining system, at the very least by altering the Equihash parameters”

blog.z.cash/zcash-evolution/

The big question is what is stopping the team from implementing a simple change that many other Equihash altcoins have successfully performed?

Zooko, in particular, doesn’t just seem to not support an ASIC resistance fork, but he actively endorses ASIC mining. The whole situation here is bizarre (if you’ve seen “Who is America?” this reminds me a lot of the mosque scene, in that Zooko is promoting the benefits of something that the community clearly does not want). I’ve come to believe that he has some sort of deal with the ASIC manufacturers. It’s the only plausible explanation for his strange behavior at this point.

3 Likes