Topic: CoinTerra announces its first ASIC - Hash-Rate greater than 500 GH/s - page 107. (Read 231016 times)

Obviously that is not the case now, but I would expected efficiencies to improve quite a lot up to that point.
And yes exactly Cointerra are saying that on their website:
But you are correct for "at the wall" consumption. I have forgotten to include the efficiency of the PSUs, where an additional ~ 10% more electricity will be used.

But even so 650000USD would be enough for around 200PH/s I believe, which again would be futile, because BTC production will only cover electricity costs.

0.55W/GH?   What using magical unicorn miners? 

Here are "at the wall" efficiencies of various devices and proposed devices. 
https://bitcointalksearch.org/topic/break-even-difficulty-by-hardware-efficiency-power-cost-value-of-btc-281279

Nobody, not even cointerra is predicting half a watt per GH/s at the wall.

My calculations say otherwise, unless they are wrong, so please correct me

So 3600 coins a day, which might increase up to 5000 coins daily with the always increasing difficulty and less than 2 weeks finding of 2016 blocks.
So 5000 coins = 650000USD daily at current market rate.

If we take a very conservative rate of 0.15USD per kWh, then 650000USD equals 27083.33USD per hour divided 0.15USD per kWh = 180555.5555555556kWh * 1000 = 180555555.55 Watts per hour

If we take an average consumption of 0.55 watt/GH/s then 180555555.55 Watts per hour = 328282828.27GH/s = 328.28PH/s

Or the other way around:

236PH/s = 236000000GH/s * 0.55 watt/h = 129800000W/h = 129800 kW/h * 0.15USD per kWh = 19470USD per hour * 24 hours = 467280USD daily.

So basically the total limit of BTC that can be mined in 24 hours is enough to cover the power costs of 328.28PH/s miners if electricity costs 0.15USD per kWh.

Hopefully that is correct.

Yes way to expensive considering the difficulty at year end, they need to offer $3 per gh/s or less to be profitable and worth buying.

It is not largish it is just beyond silly.   It think you are misunderstanding.  236 PH/s is so large even if the hardware was free you would mine at a loss because even with @ 2J/GH and $0.10 per kWh it would cost ~$300 per Bitcoin just in electricity.  So if someone gave you a free miner you would lose money by using it.  When you consider that even @ Cointerra prices (lowest price per GH) we are talking nearly a $1B in pre-orders (or 8 years of gross mining revenue excluding electricity) it just gets even more silly.

Difficulty is going up but saying 236 PH/s is plausible well you might as well say difficulty of eleventy quadrillion septillion is also possible.

Sorry.. I didn't mean December, I meant July 2014... http://mining.thegenesisblock.com/a/7a0c452d0f

I have been thinking one thing and typing another, sorry for the confusion.

"Perhaps"? No way! There's no way the bitcoin network is going to approach that kind of performance by the end of this year.

I agree perhaps 236 Ph is a large estimate for the network speed in December 2013, but I think with all of the new companies coming onto the scene it is not impossible. Cointerra, Labcoin, KNC, HashFast, BFL, Avalon, VMC, ASICMINER... all designing and producing their own chips and all scheduled for Q3-Q4 shipping dates. There will be other companies that spring up that buy chips from these companies and produce their own equipment as well.

I also agree with you in that no one will buy mining equipment if it is unprofitable to do so. However, until that point is reached I feel like people will continue to throw money at hardware investments and pre-orders until then. Mining manufacturers will lower their prices over time until that point is reached (rather quickly) where it is unprofitable to buy consumer mining equipment. Difficulty increases will not stop then, there will be many backdoor mining operations setup by these companies as long as it is profitable for them to do so. It will be profitable for them to do so long after the point it is unprofitable for the consumers.

We're already seeing that with 130nm Bitcoin ASICs. A company in China can't make any profit selling the equipment to international customers because the ROI for the customer would be horrible when you include shipping costs: https://bitcointalksearch.org/topic/tmr-asic-project-introduction-281866

However, just because it is unprofitable to sell them internationally, it doesn't mean it is unprofitable for them to make a large batch for themselves and mine at exact manufacturing costs of the equipment. It will take a much larger difficulty increase for this to happen. This is an unforeseen factor that I think most miners are not accounting for in the increase of hash power. The manufacturers will be competing with their customers. I don't care how honest/ethical you are, if you have an opportunity to make Billions/Millions of dollars by doing something that it slightly unethical... as they say, money makes the world go round.

cointerra reduces it's prices on the 2th/s miner to 13,999.  According to the mining dashboard, you'll earn a negative $1,060 on that deal if you start mining on december 1. if you start january 1, owing to late december arrival (and they aint promising better), then you're totally jacked at negative $7,000.  Common guys throw us bone. We really do want in on second generation asic but we ain't that crazy.  Maybe they're banking on the fact that some of us are actually pretty crazy.

No it isn't realistic in any fashion.  236 PH/s would mean that every single miner in the world would be operating at a loss.  So even when miners have a 100% negative ROI from day one people will keep deploying tens of thousands of more units.  Why?   They have to much money and want to turn $1000 in electricity into $50 in BTC?

This seems a good point to bring up something that has been puzzling me for some time.

In the manufacture of CPUs and GPUs there are always chips which don't quite make the grade, and so are binned and used to make products with lower specs (frequency, number of cores etc.), which leads to families of products with different price points. But none of the Bitcoin ASIC manufacturers are following this business model. Why do they expect to get close to 100% yields with their chip manufacture? Is there something about SHA-256 hashing that means the chip design to manufacture is more robust? or are the companies wasting the lower quality chips? I think there are plenty of people who would consider buying a less productive ASIC if the price was right, regardless of the manufacturer.

Yes, it seems reasonable considering this: http://bitcoin.sipa.be

There are so many companies and securities releasing hash power in Q4 this year. I would not be surprised if the network continues its current growth rate through the end of the year. Also, I think a large percentage of manufacturers will setup large mining farms using their own equipment at cost. Notice how none of the newer companies have stated they will not mine on the Bitcoin network. This will push difficulty up much higher & much faster than anyone anticipated.

This is not me hating on just Cointerra by the way. I think almost all current consumer mining hardware equipment is overpriced. They are not the only ones by far... I just would like to see them succeed because they are from my home state of Texas, so I am giving them constructive criticism.

To lose $1200 difficulty must be over 33 billion, which equals 236PH/s network speed.
That speed equals 472000 (472 thousands) 500GH/s miners or 118000 (118 thousands) 2TH/s miners.
Does this sound realistic to you?

I'm glad they have taken my advice & advice from the community on lowering prices.

However, it's still not a good investment. It seems they're trying to find the highest price that the market will bear. I don't think they're there yet, but certainly some people that don't understand the economics of Bitcoin mining will still order from them.

Even if delivered on time you're likely to lose $1,200... not exactly my idea of a good investment. http://mining.thegenesisblock.com/a/85e10d5a76

They don't seem to understand the risk/reward ratio needs to be just right for people to risk this much money on pre-orders. If there is no possible reward even if they deliver on time, I don't know how they expect to sell their products (unless they are just planning on mining with the equipment themselves at cost).

Please, no more pre-announcements, we want to see the production working model ready to be purchased and ship out.

but nearly completely content-free.

All the math in the opening sequence has nothing to do with Bitcoin, its some partial differential equations like used in financial derivatives. At one point there is a close-up of cubicle with wall calendar open on August 2013 and close up of a desktop screen with some Verilog code and some digital simulator. If they are still working at the Verilog level this month then they are either very early in the design or going to completely offload the design to Open Silicon.

Also hiring a cryptographer is completely technically pointless for miner. They should be working on the analog/thermal/physical design phases to deliver anything meaningfull for long-term Bitcoin mining.

http://www.coindesk.com/cointerra-cuts-price-of-terraminer-iv-bitcoin-mining-rig/

chip photo is ivy bridge with cointerra logo shopped in, video is nice though
http://vimeo.com/73209233

#4 is interesting.  It would logically seem their only options are hashing power in multiples of 500 GH/s as that is the capability of a single chip. 

I might be interested if there was a contract with a black and white hard deadline date and penalty of 25% extra hashing power per week (exactly 7 days late) up to +100%.

Something like
2.0 TH/s if shipped by 31 DEC JAN 2014 or earlier.
2.5 Th/s if shipped before 7 JAN 2014.
3.0 TH/s if sipped  on 14 JAN 2014.
3.5 TH/s if shipped after 22 JAN 2014.
4.0 TH/s if shipped after 28 JAN 2014.

Note for the reading impaired this is NOT Cointerra's protection plan just IMHO what it will take for such a later offering.


Still it does make you wonder what BFL Monarch buyers are thinking.  Cointerra has at least BFL beat in every possible category even if we assume "this time will be different" for BFL.
                           

BFL Monarch is higher cost per GH ($8 per GH vs $6.95 per GH)
BFL is lower efficiency (when you consider entire system "at the wall"  (https://bitcointalksearch.org/topic/m.3008390)
BFL is promising a later delivery (February after the whole rocket run already sold out bait and switch)

This is also assuming BFL will actually delivery on time, on spec so it could be considered a best case scenario for BFL.    Why anyone bought a Monarch I have no idea.

These guys are still bugging so hard:
1. Price is still too high for December Delivery
2. Offering cheaper, lower end units fixes nothing. We're not scoffing at the price tag, we're scoffing at the value.
3. Giving a 20% bump in hashing power for a 30 day late delivery calms no one given that the difficulty is growing at a rate of 70+% a month
4. How do you even offer a 20% bump in hashing power when each chip represents 25% of the unit's hashing power?

Bitfury "ship next week" 400ghs miner resales for usd 4000

usd 4000 x 5 = usd 20000 is the price for 2trh now, for 55nm devices.

with difficulty projected to be 5 time up in January, dividing 20000/5 = usd 4000 for 2trh.

Most miners investing mined coins into new equipment. By December people may not have money to pay even 4000. Because, it is possible that 28nm devices will be available as ship now from Avalon and ASICMINER by November, BFL too ... hell yeah, and all money will goes there.

Cointerra have to be real much more aggressive, if they want to grab some attention.