Topic: Bitfury 16 nm chips and board design - page 3. (Read 14851 times)

Thank you Bitfury, sign me up sidehack!

I'm definitely loving this and I'll be in for whatever comes of it.

good news as always I have 2 coins set aside for you when needed.

And I would love to run a new club using bitfury gear.

Good news sidehack.

Great news!  Can wait to see that. Will for sure be Canadian Supplier with pleasure. 

It's not my job to say anything officially, but someone higher up the food chain has been working and it looks like Bitfury actually likes my project idea so it's looking more possible now that it'll actually happen. Got a lot of design and prototyping to do before anything's for sure, but it looks more possible. So that's pretty exciting. Don't ask me for details on the politics of anything, because I don't know. Like I said, someone higher up the food chain's doing the talking.

punin was talking on some other thread that they will be allocating some chips to sidehack, which is good.

found the reference:
https://bitcointalksearch.org/topic/m.13831108

since price is unknown and it was a bottleneck, i wonder if something got cleared recently.
i also might be interested in chips buy/miners, but in modest numbers-would be nice to know if there will be a threshold of participation.

Count me in for a certain number.

Interested in this!

I'm talking to some guys who could be on the A Team for design, since one person really can't do everything quickly or necessarily as best as it could be. As Phil's pointed out somewhere, with the crazy diff spikes and halving approaching, even a super-efficient miner might not break even unless the chips are not super expensive. The viability of the project depends at least as much on the actual chip price as it does on the ability to raise enough money for MOQ.

Just want to say that I would be interested in a chip group buy if sidehack could make a miner with them.
I'm also interested in your hosting service, sidehack.

I am going to throw this out there and see if it sinks.  I am all in because there are a lot of wheels in motion.  Block Reward Splitting, S7 nearing end of life in production, possibly no new hardware between now and April and the biggest part of it is that the price of BTC is stable at BTC

Look at this from Bitfury's standpoint.  The have a lot of BIG NAME investors looking at them.  They need to be profitable and that will not happen with the numbers above.  How many multi million dollar Immersion Data Centers that require 1.5+ megawatts are Bitfury going to sell if the price remains the same?  Not many, hence I am hedging all my efforts that the powers that be will make sure they are profitable. 

If I had the kind of money and backing that BitFury has it would not be difficult to get people excited about their stuff if the price of Bitcoin just "happens to shoot up" around the end of March.

Except that's now looking like about a 72-chip miner, so to make it for $600 would require chips under $6 or so.

Also, there's a mistype in my previous post - 4TH 480W not 240W.

Also if I'm thinking right, something 0.3J/GH at 6c power will still turn a profit with $400 coin and 25BTC blocks up to a diff of about 460B, which we could see before the halving. Something 0.3J/GH at 2c power and $400 coin is still turning a profit up to about 700B after the halving? That sucks. I wonder how much gear will be turned off at the halving. Think the diff will drop?

Something 0.12J/GH at 6c power and $400 coin should still be turning a profit at diff about 580B after the halving. By then taking it down to 0.07J/GH keeps it in the black all the way to 1T diff?

I will play with that set up .

If you can build 4 th  at 600 usd

that uses .07 watts  which is 280 watts  it could do roi right near the ½ ing

based on  6 cent power----------
based on 600 usd price for the miner

based on 399 usd a coin.

and based on 7% growth.

the need for gear may be not for filled   as diff is that high.




I believe s-7's built at 300-400 usd mining at 2 cents  are doing the diff jump

What's the numbers for a 48-chip refit giving you 4TH 240W wall down to 2.6TH 200W wall for, say, $600 all in? That could be possible with ASICs at about $8, which is really about the most I'd want to pay.

so lets look at growth and when would chips get into sidehack's hands

crunch numbers yada yada yada this is a bust at 1,000,000 usd for 75,000 chips.

even if a board was made and ready to populate right now.

lets say current hash growth has been done by bitmaintech to hurt the bitfury chip.

they succeeded.

if I make a bitfury miner that can do .07 watts to .12 watts   per gh  20 chip miner

1th to 2th   70watts to 240 watts  the 20 chips will cost me 260 usd

 controller software etc I can't build a board to fit an s-3 as a drop in fit for 300 usd not with chips costing 260
the assembled  board and the shipping  will be at least 350 usd (certainly more)



So I tell an s-1 or an s-3 owner give me 500 for 1 drop in board   that maxes at 240 watts and 2 th

or mins at 70 watts and 1 th  .

I can mail it to you by the 175 diff mark which is in a few days.   so do an roi on a 500$ board  assuming you have an s-1 or an s-3 to drop it into.

roi at 20% is never
roi  at 15% is never
roi at 10% is  never

drop price to 400
power at 6 cents

roi at 20% is never
roi at 15% is never
roi at 10% is never

so a 2th 240 watt drop in board right now does not roi and I am not counting the ½ ing

even if growth drops to 5% it does not roi.

We are watching a play much like Gordian knot

https://en.wikipedia.org/wiki/Gordian_Knot

Chinese mining farms at 2 cent power have decided to max hash and take as many of the coins they can before the ½ ing.

A company that builds the s-7 for 200 or 300 usd and makes 1000ph of them can just add them and keep growth rate over 10%  this unties the knot with brute force.

The 14-16 gear  will lose money at 6 cents power and 400 usd for 2th in hand right now

I'd like to see more like $7-8 per chip. $13 per is gonna be hard to justify when the diff's where it is, and only going to get worse during dev time and manufacture. Ten months ago $1M would have bought about 8PH worth of BM1384 and now $1M will buy about 8PH of these, so I guess that's probably not too bad, but since then the diff's gone up 2.93 times since then, and the comparative efficiency of the two chips is about 3.4 times given the operating points mentioned. So if the diff goes up another about 18% from where it is now, these chips will be worth less now than the same dollars worth of BM1384 were ten months ago (at those given setpoints). Mind you the efficiency ratio between the given setpoint and bottom clock for the BM1384 is about 1.6 versus about 1.9 for the Bitfury chip, which means BF should have a longer viable lifetime if undervolting is possible. But you also have to consider that 10 months ago the BM1384 had been shipping for over 3 months, rather than negative two months.

It was 1M for a full lot of ~75k chips. That already has the discount Not sure they can realistically maintain this price with current diffs and halving soon. Even if we optimistically get a working miner by April, bit fury will have already added their 16nm hash to the networking by then which is at least another 500 PH. We would be looking at a diff of about 200 billion by then, which translates to a profit per chip per month of about $2 for 3-4 months before halving.

Chip would have to be bought at under $10 per to be profitable. (of course this does not count costs for actually producing a board/miner etc).

From what I see declared interest for 35000 chips so roughly 350 k
May be you should think about starting a tread for the people just to express interest in terms of quantities/USD.
The leaked chip price is 13 USD and I think it will be better with 1M order

Wait. Hang on. How many people are looking at me to head this up?