[Announcement] Avalon ASIC Development Status [Batch #1]

[PuertoLibre]

Assuming bASIC and BFL ship around the start of December.

By the end of Decemeber, I expect bASIC to have shipped all units .

...it will require at least a 7 days of 8-10 hour days,

~800 * 54 = 43 TH/s (assuming bASIC performance stays at 54GH/s..it will most likely be higher)

BFL has more orders and more man power...let's be really conservative and say they ship 75GH/s in Dec.

thats 43 + 75 + 22 = 140 TH/S bare minimum

So that IS ~7X difficulty from day ONE for Avalon customers. I believe your "conservative" ramp up numbers are not conservative enough.

Regardless, I'd say a very "conservative" guess is that within 3 months of Avalon's release, difficulty will be over 10X (IMO, easily). Long terms calculations should begin a minimum of 10X.

Everyone has been saying time-to-ship is the most important factor...well I'd like to look slightly further out than 3 months. The most that I can conservatively plan for is 6-10 months. Who knows what the landscape will be like several months from now.

My outlook for the short-medium term is 10x-30x difficulty. All 3 major offerings will be profitable during this time...how profitable depends on power cost.

Since BFL has published (confident) power numbers...let's use them as a baseline.

Cost per day @ $0.10 @ 60W = $0.14
Cost per day @ $0.10 @ 100W = $0.29
Cost per day @ $0.10 @ 200W = $0.48
Cost per day @ $0.10 @ 400W = $0.96

Cost per day @ $0.25 @ 60W = $0.36
Cost per day @ $0.25 @ 100W = $0.60
Cost per day @ $0.25 @ 200W = $1.20
Cost per day @ $0.25 @ 400W = $2.40

25 BTC/block @ 30M difficulty @ $12 USD/BTC = $12.07 / day

Let's look at how much profit is effected by power

using 60W as baseline
Cost per day @ $0.10 @ 60W = baseline is 0% more profitable
Cost per day @ $0.10 @ 100W = baseline is 1.27% more profitable
Cost per day @ $0.10 @ 200W = baseline is 2.94% more profitable
Cost per day @ $0.10 @ 400W = baseline is 7.38% more profitable

Cost per day @ $0.25 @ 60W = baseline is 0% more profitable
Cost per day @ $0.25 @ 100W = baseline is 2.09% more profitable
Cost per day @ $0.25 @ 200W = baseline is 7.72% more profitable
Cost per day @ $0.25 @ 400W = baseline is 21.10% more profitable

Obviously, as difficulty goes up, profitability goes up. As such, the affect of power usage on profitable also goes up.

Some people might say 5% is nothing...we tip pool operators that much. Others are trying to maximize every penny. Granted, Avalon now claims a 10% edge in hashing power (over BFL)...I am uncertain how long that edge will last. Both bASIC and BFL have stated that there is (perhaps significant) room for improvement here.


I have orders from each of the big 3, so I'm not trying to bash Avalon here. As a customer I just want a great product that will give me a decent bang for my buck.

BFL offers good hash rate with lower power.
bASIC offers good $/GH...still waiting for power numbers.
Avalon still has good $/GH (with bumped specs)....still waiting for power numbers.

I tried to diversify for the first run of ASICs, but if/when I decide to order more units...I'm looking for the best bang for the buck going forward.

I hope you keep us posted on your impressions of each device and it's performance/watt.

It is rare to find someone whom can compare all three in the same locality. Your reviews will be indispensable.

[arklan]

this whole power debate is exactly why i have always planned to set up solar/wind/whatever power sources asap for any significant mining effort. ...i should set up a solar selling business for BTC...

[PuertoLibre]

this whole power debate is exactly why i have always planned to set up solar/wind/whatever power sources asap for any significant mining effort. ...i should set up a solar selling business for BTC...

Rather than solar alone you should look into companies that are pioneering technology to convert heat (on your roof) into power. Most average sized houses have a 47 kilowatt hotplate right above their heads.

All that energy goes to waste every day. Companies that are working to harness that thermal energy are probably going to be very rich in the future.

Some thermal waters heaters (not the solar panel type) for example, require very little roof area to heat water. Converting the rest of it into actual electrical energy is the sweet spot in alternative energy.

[Zeek_W]

this whole power debate is exactly why i have always planned to set up solar/wind/whatever power sources asap for any significant mining effort. ...i should set up a solar selling business for BTC...

I have wind, hydro and solar setup on my property. I don't live there but when I do - free everything!

[pieppiep]

this whole power debate is exactly why i have always planned to set up solar/wind/whatever power sources asap for any significant mining effort. ...i should set up a solar selling business for BTC...

I have wind, hydro and solar setup on my property. I don't live there but when I do - free everything!

If I want 'free' energy like that, I still have to buy the turbines/panels, so there still is no free energy.

[Coinoisseur]

What do you think about this?
Especially as regards the European market, with electricity costs between $ 0.10-0.30/kwh

strictly speaking on power being the only variable I wouldn't even suggest to "mine-for-profit" basis with power rates higher than $0.10 ( and you are unable to get write offs or reductions. )

Sorry, but you know, in how many countries in the world the power rates is cheaper than $0.10, right?
In Europe I know none! This means, that you throw away in minimum the whole European market for  "mine-for-profit" or "long term mining as an investment" customers!

The main inputs of mining are equipment cost and electricity cost, decent mining setups require pretty light amount of man hours to maintain. So like other electricity sensitive activities areas with the cheapest power will have a distinct advantage in terms of ROI rates. Unless Europe develops more efficient hardware and then restricts export of said hardware it's going to be at a disadvantage in these activities given current electricity costs.

[DoomDumas]

I keep reading about estimate, prediction, comparison.. power usage, difficulty, Gh/s...

Everytimes, ONE BIG THING comes to my mind...

Is'nt all a big guess about the future "price" of a bitcoin ?  

No one can predict the price, I assume that's why all calculation and estimate are done on a 12$ basis..  Ok, we need a base to make up our mind..

BUT, it almost a certitude that the ratio of exchange will not be 12 USD/BTC in two years, even in few month !

As we saw in the last 2 years, the ratio of exchange USD/BTC have been between 2 and 30  (let's say 32 to simplify)  from 2 to 32 it's 4 times doubling or dividing.

Some scenario on a 2 year time span :


(IMHO) realistic possibility : X4 or /4

12 double 4 times = 192 USD/BTC ratio.  peoples will then say, OMG, why I have not ordered tons of those asic !
12 halved 4 times = 1.5 USD/BTC ratio.  I dont want to write what peoples might say !
major event : 0 USD/BTC  (it's not impossible)


(IMHO) conservative possibility : X2 or /2

12 double 1 time = 24 USD/BTC : enought to say I would mine with a power cost of 0.30$/Kwh
12 divided 1 time = 6 USD/BTC : hard time for miner !
major event : 0 USD/BTC  (it's not impossible)

As an optimistic : I'll give it a try... I bet on a 100 Fiat/BTC ratio within 2 years....

Buying asic..

And I'll never invest more than I can afford to lose overnight because "major event : 0 USD/BTC  (it's not impossible)"


Was my 2 satoshi

[Syke]

Is'nt all a big guess about the future "price" of a bitcoin ? 

So take price out of the equation. A slightly better way to estimate future returns is on BTC output verses BTC cost today.

Today's price of $12/btc, $1299 = 108 BTC for an Avalon ASIC.
66 GH/s, 12.5 btc/block reward

That leaves only 1 variable, network difficulty. ASIC pre-orders are going to hit 10x pretty quickly, beyond that it'll probably taper off.

10x network difficulty = 1.08 btc/day

100 days to recover spent coins. Funny how that turned out to be a nice round number.

[Transisto]

About power figures,

If the same energy now spent on GPU get spent on 1w/ghs ASICs the difficulty will be 450x what it is now or ~1.38 billion.

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

[mrb]

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

Are you one of these guys who think that "640kB ought to be enough for anyone"?

If Bitcoin continue to succeed, it will come to a point where $200M will have been spent on mining hardware.
Just consider that today alone, $20M has been spent on CPU/GPU/FGPA mining hardware.
Going to $200M is only a 10x increase.

[pieppiep]

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

Are you one of these guys who think that "640kB ought to be enough for anyone"?

If Bitcoin continue to succeed, it will come to a point where $200M will have been spent on mining hardware.
Just consider that today alone, $20M has been spent on CPU/GPU/FGPA mining hardware.
Going to $200M is only a 10x increase.

Unless by then $ stops to exists and you buy your hardware with BTC 

[Transisto]

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

Are you one of these guys who think that "640kB ought to be enough for anyone"?

If Bitcoin continue to succeed, it will come to a point where $200M will have been spent on mining hardware.
Just consider that today alone, $20M has been spent on CPU/GPU/FGPA mining hardware.
Going to $200M is only a 10x increase.

The 20m$ GPUs pay themselves in ~300 days with current profit/watts, while the 200m$ = ~20 years (Edit: Actually never given reward halving)

The price would need to be of an average of 240$ during the next four year for hardware to pay for itself in the same duration.

In other word, very plausible given the imbalance in monetary distribution, Unfortunately my planing for a such a scenario is right next to my alien invasion preparedness folder.

[Xfinity]

Is'nt all a big guess about the future "price" of a bitcoin ?  

+1, we can only speculate at this point. We will soon know ...

[makomk]

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

That would be 200M $ worth of BFL hardware at current prices. There's a substantial financial incentive for them to drop their prices once business slows down - the marginal cost of ASICs is fairly low and the up-front cost fairly high, and they probably want to make as much money on that upfront investment as possible. Of course, that doesn't leave much room for purchasers of BFL's devices to cover their investments either.

[ercolinux]

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

That would be 200M $ worth of BFL hardware at current prices. There's a substantial financial incentive for them to drop their prices once business slows down - the marginal cost of ASICs is fairly low and the up-front cost fairly high, and they probably want to make as much money on that upfront investment as possible. Of course, that doesn't leave much room for purchasers of BFL's devices to cover their investments either.

Indeed pushing too much power in the network in a time brings to very low profit unless price of btc skyrockets. So if ASIC producer don't want to kill themself they have to don't reduce price or put too many ASIC out. Doing some small calculation if network power jumps at 40X a 60GH/s ASIC can barely cover the energy costs in USA and noway is profitable in lot of European country (in Italy you've to add 500$ in a year to cover energy cost).
So or price of BTC double in next few months or power increment due to ASIC introduction will not be over a 20-25X or there'll be no profit in mining

[Aseras]

Something many of you are forgetting is that the bitcoin infrastructure as it is now, with pools and current hardware will essentially go bye bye once the asics start hitting hard. It's going to turn the world upside down for a few weeks/months. There's something like 20t/hash now and a majority of that is going to evaporate a month or two into asics.

[pieppiep]

Why would the pools disappear?
If someone mines with a single gpu now and has 1/10000th of the total networkspeed and gets a asic and still has a 1/10000th of the total networkspeed, it's still a good idea to use a pool.

[ercolinux]

Why would the pools disappear?
If someone mines with a single gpu now and has 1/10000th of the total networkspeed and gets a asic and still has a 1/10000th of the total networkspeed, it's still a good idea to use a pool.

A mining rig of 500MH/s today have 1/40000 of network power a single 60GH ASIC at the beginning will be near 1/2000 (like a 2GH today) and a 1TB rig will be in the 1/100th range like a small pool today (30-40GHs of today rigs); pool still can work, but it's quite easy for who have 2-3 60GH ASIC mine in solo and get 1-2 block a month in the first month.

[dani]

Is there anything planned after avalon? Avalon II maybe? Or maybe just another batch? Didnt read anything about it, nor from any other competitor.. just curious

[MrTeal]

Packaged Chip Size: 7 mm x 7 mm

Thanks for the transparency, but the unpackaged die size is what you ought to be posting.  Without that it's impossible to make performance comparisons (or at least those comparisons would have to assume it fills the whole package cavity, which portrays Avalon in the worst possible light).  Please provide the unpackaged die size.

Also, I think you may have neglected to say how many of these chips are in the 60GH/s device… or perhaps I missed that.  Anyways, without knowing how many chips are in the device, none of the information above is really useful.  Did you post this somewhere that I didn't see?

Assuming a maximum die size of about 25mm^2 in a 7mm QFN, I would guess quite a lot of them if the unit is actually drawing 400W. Probably a few dozen. That's a lot of heat to move from a molded package.

That would be 200m$ worth of BFL hardware... Not going to happen, leaving quite a margin for less efficient device to cover their investments ... at some point.

That would be 200M $ worth of BFL hardware at current prices. There's a substantial financial incentive for them to drop their prices once business slows down - the marginal cost of ASICs is fairly low and the up-front cost fairly high, and they probably want to make as much money on that upfront investment as possible. Of course, that doesn't leave much room for purchasers of BFL's devices to cover their investments either.

There's definitely a lot of room for BFL to come down, but I'm not sure just how much. I could see the Minirig SC come in at $15k, but it doesn't matter if the ASICs fall out of the sky, you're not going to see a Minirig SC at $1500 until there's a respin that brings down the power and size of a 1.5TH/s machine to close to the size of a Single. My speculation is that BFL could sell the Minirig SC quite profitably at the $15k price point of the current MR, but it would not be economical at $7500 (1/4 it's current price) unless they start to produce tens of thousands of them.