Topic: Mining Equipment Manufacturers - page 14. (Read 55732 times)

First I've seen of them actually.  Looks like they have been taking orders for over a month at least, though, and were supposed to deliver a couple days ago.  Maybe this is a typo on the page, since it says they have 2500/2500 machines.

Any thoughts?

http://www.german-bitcoin-mining.de

Andreas indeed chimed in, no big surprise there.  : )   But prior to the discussion they played an audio quote from Dan Komensky (sp?) at Bitcoin San Diego 2013 where he made a claim that the current proof of work algorithm would not last another 12 months. He's of course not the boss of Bitcoin, but the boldness of his statement and the implications of it got my attention.

Thanks.  And...  how funny, I wrote an article on the 500GH, I guess I checked it off in my head and forgot to add it in.


nightengale, I haven't heard the episode yet, I'll get back to you   Sounds like something Andreas would say.

Please forgive the off-topic post, but I'm throwing this out here because I figure a lot of current and future ASIC users follow this thread...

Has anyone ever considered the possibility that the Bitcoin foundation could change the proof of work algorithm to scrypt or something else? I've got a finger in a few pending ASIC offerings, and I was listening to Let's Talk About Bitcoin today and I realized how disruptive such a move would be for all of us. Had anyone read anything about the likelihood of such a change? As if the hash rate and price of BTC weren't enough to worry about... 

You can't buy avalon any more they only sell chips so should not be on his table
not sure why BFL 500 GH is missing

This table is very usefull information and I like to thank   FCTaiChi

and the ROI can be used as guide as no one hnows what the hashrate will be in 12 months time

We might be approaching it.  Everyone has to be working on the more efficient chips at this point.  There is a lot of room for growth in the ASIC market, we haven't even maxed out the design potential yet, and most companies are still working with the first thing they came up with when everyone realized ASICs were the way to go.  I really don't know how long these 20% gains will last, just saying it could be a while.

I'll probably mess with the range just a little more before it looks right, there are so many possibilities..  the price is horrid right now, we could have a mass exodus away from mining if it goes too low, or the opposite could happen.  We'll see what happens with MTGox regulation and technology.

I haven't been able to go through the order page for over a month now.  I don't see where they are available to order from.  Actually I used to have this model on the table, but for a while it said out of stock next to it.  If you can't order a miner it doesn't go on the chart.  This is for people who are looking to order. 

Very soon, though I will have a section up for those who ordered to see how much they will get in return too.

If you mean about holding the chip orders out until November, I would love to change that, it is over double what the company says.  They have given very strange explanations (that everyone else seems to accept just fine) and if I were to push the date out to what it should rightfully be just by the equation:
Average lead time + today's date
I would have to put it out much further than November.

I want you to know that I have no problem with your table. It's a great tool to find out different possible scenarios. Just wanted to point out what a constant diff increase means to total network hashrate. And I also agree that in next month we will see increases in the 20% range, so you can leave that calculations on table too. But it will slow down - current generation ASICs are still rare, and not on time. And overpriced for a hobby miners. Next generation may increase the hashrate growth again, but we are not there yet.

Jadatking must be stupid, because you pay more for one at hand. That one without receiving.
https://bitcointalk.org/index.php?topic=245629.0;all
Jadatking is not stupid is a figure of speech.


if you want a bit on the forum you can read. more users without their avalonbatch2.
from my point of view I think that just trying to help the community.

Avalon Batch 2 finished shipping I think FCTaiChi hate's Avalon for some reason

 Stilll no entry for the avalon on your chart?

They are shipping you know, more so than BFL.

The reason I had the higher percentages posted was because I thought that having a greater portion of rise in the beginning of the period would effect ROI.  I just went and did the math and I was wrong.  I see no need to keep them there, and I agree with what you guys are saying, I just misunderstood the math.  In fact I think I've said before that I think any estimate over 17% would take a completely different situation than what we see now.  Not impossible, just a lot of new tech needs to come out. 

I changed it to a 8%-16% spread, does that look better?  It does seem a little weighted toward high difficulty, but going any lower seems ridiculous too.

 
While I slowly wrote that short post above (with the help of google translate), Schrankwand already explained all that.

Don't look just at the averages, also look at what was happening at that time: network expansion, more and more miners coming to mining, then GPU hashing. And GPUs were readily available in every computer shop, not some pre-order BS that's happening now. And even that GPU growth (average 30% EACH retarget) stopped - in about a year (jul.2010-jul.2011).

The log chart at http://bitcoin.sipa.be/speed-ever.png

shows that ASIC technology doesn't have such influence as GPUs back then (the curve is no so steep). That's because ASICs aren't readily available to everyone (pre-order game), and also because network is not so 'young' now. Here I mean that you need greater effort (more hashes, more money invested) to increase the network hash power by the same percent as before with GPUs.

As network grows, that effort also needs to get bigger, if you want to sustain the same pecent of growth rate.
Or, there must be lots and lots of new hashing products - cheap, powerful, and readily available to every miner.

I agree, and say quite often that we should expect a plateau soon.  However, it doesn't seem plausible that we will look back after a year and see a difficulty rise of less than 10%.  
This is based mostly on looking at historic data, which there is only a few years of, and we have never seen a change as dramatic as the ASIC craze.  I'm quite open to the possibility that the data in the table does not encompass every possible scenario.  The "constant increase" is definitely not a model for reality, but anything is a guess.  I'm adding more information every day to allow for different interpretations.

This is just a start, but for instance every manufacturer will get their own page.  The page isn't finished, but here is preview.  What would you think would make this more accurate?

http://decentralizedhashing.com/bitcoin-mining-equipment-table/kncminer/

Yes, of course. However, this is to be expected in a system that is based on revolutions. I propose that from previous system patterns, it is likely that we will see a drop off of increases based on price.


Here are some examples, I would like to give:

In the beginning, people used CPUs. Then people used GPUs. FPGAs had similar hashrates, lower power needs, now come ASICs.

Look at the hashrates. You will find, that this is the graph:




NOw, this is historical. You will see that at the first revolution point, somewhere around May 2011, you saw a HUGE increase from GPU switching.

It was unbelievable and you might expect that people would just keep adding and adding and adding and IT DID NOT HAPPEN.


Instead, what happened was what ALWAYS happens in system theory: A new systemwide equilibrium was reached. The next revolution comes now, January 2013 and it just keeps on going.

But the price does not keep up. And this is quite important: The cost of adding new hardware, however trivial it might be, will be always weighed against operating costs and break even on investment.

As you can see, we can take a look at what happens here:



April 12 until January 2013. We see an almost doubling in price and what do we see on the miner charts? Well, we see that the price seems to be moving along. Or do we?

Either the price goes up because the good is getting more scarce, THAT would be in fact classic economics. I have studied my fair share of psychology and therefore say that this is likely hogwash! What is more likely is that miners are increasing their hashing power in response to possible rewards. Greediness is most often the better explanation.

I would suggest reading the recent paper here:
How psychological framing affects economic market prices in the lab and field Ulrich Sonnemann, Colin F. Camerer, Craig R. Fox, and Thomas Langer. . PNAS, 2013 DOI: 10.1073/pnas.1206326110
http://www.pnas.org/content/early/2013/06/26/1206326110

So, what does all this have to do with the discussion before? It is heavily unlikely to get such an increase in hasing power for the reason that this increase is absolutely idiotic if the price doesn't go up. The good becomes more scarce, people will not try skimping for tiny sums, that is how humans are.


COming back to the hashrate: If we have a hashrate increase to 4,5 Petahash/s, the difficulty will adjust. It will adjust the same 23,99 i have mentioned before. That is an increase to 511.834.542.

At this difficulty, a 50gh/s Single from BFL will net you a whopping 1100 per year at current price. It will get to a break even of over 2 years.

And this is why psychology will be important: People will stop investing. They will get shitass scared because the "fast riches" are going to stop being delivered long before that. LONG before that.

You're close, but not quite on target. 
If you look back through the history of bitcoin the average total increase per period is 17.8% or 20.24% if you take out the year when growth was 0%.  The four month average is 18%.
Depending on what the companies do anything could still happen.  There is a lot of competition happening out there.  Most people do agree that the miners that are being made right now have a very short time frame of return, so this will have to drop off pretty soon.  However even if we do only grow 9% in the following year we could have a ROI that looks more like 25% if the growth is weighted toward the beginning.

Looking at the difficulty rise it's hard to imagine that it would go that high.  To follow the trend that has been set since the beginning of bitcoin, though, we will have to see some pretty high numbers. 

Yes, but it still assumes  a constant increase. A constant percentage increase in difficulty would mean that every time you do this, you add 10% to the added hashrate before.

To make things clear, here again. One year means a theoretical retargeting every 14 days.

One year is 52 weeks, 14 days is 26 retargets.

So, for the smallest increase, we need 1,09^26 which comes to: 9,39
13% comes to a whopping 23,99
17% comes to 59,27
21% to an unbelievable 142,04
25% racks it up with 330,87


So, what does this mean?

This means, that if we assume, from current difficulty, that this happens, these hashrates have to be achieved:

9%  --> 189TH/s times 9,39 --> 1774TH/S aka 1,77 Petahash/s
13% --> 189TH/S times 23,99 --> 4534 TH/s  aka 4,5 Petahash/s
17% --> 189TH/S times 59,27 --> 11202TH/S aka 11,2 Petahash/s
21% --> 189TH/S times 142,04 ---> 26845TH/s aka 26,85 Petahash
25% --> 189TH/S times 330,87 --> 62534TH/s aka 65,5 Petahash per second.


I mean guys. Are you serious in believing most of these? I want to be honest here. I believe, we might, and this is more a worst case, we MIGHT scratch the 1,77 petahash mark within 12 months. And this is with a lot of BFL and Bitfury delivering in worldwide fashion.

4,5 petahash within one year is already a dream, if you ask me. But, I will not say it is impossible, since we live in interesting times engineering wise.

But anything beyond that is ludicrous. You people are able to write a chart and create mathematical formulas. You need to be able to understand that increases in difficulty build always on the former difficulty. A 50% increase today is a corresponding 33% increase in the next retarget if it comes with the same hashes. That means it becomes exponentially harder to change the difficulty target.

This is what Bitcoin has always been about: Creating an ultra stable supply supported by early adopters.


But if you guys think that 11petahash are alright, I would probably sell all Bitcoin and do other stuff. Because we are still in Gold rush mode, everyone. And the moment where only bigger mining corps (Like in today's gold) win, is not here yet.

And 21 and 25% constant retargets? Do I even have to talk about that? Did you READ those numbers? Did you even make the calculations of money that would have to go into this system?


Guys, this is money. And yes, I understand that with money, people tend to shut their brains off. But a one year 25% constant retarget is unbelievably ridiculous, that not even BFL Josh in his wettest dreams could think of the money needed to do that. Let me clarify it with bitfury chips: If one Bitfury chip has 5gh/s, we would need 12 million, 650 thousand, 800 chips, in one number 12 650  800 chips to reach this number. The Avalon Chips we have seen have not even hit the million mark yet. This will never ever happen in one year.

The number you guys want to watch is the 9%. This is the interesting one for next year. Anything else is incredibly unlikely.

starting to be a lot of "red" in your charts !!!   and any slipage make things worst -
october will be an intesting month