We’ve got a Butterfly Labs Bitcoin miner, and it’s pretty darn fast

The loud little box busily decodes money at 5.35 billion hashes per second.

Read the whole story

 

[tutis]

Could this kind of little box be used to brute-force password cracking attempt? Isn't it doing the same kind of math?

 

[zneak]

I'll be curious to see how much bitcoins have to be worth to make up for the electricity consumption.

 

[caw]

Is the outlet voltage in your house so bad that it was necessary to report that power AND current are constant? My condolences to all your appliances

 

[DerHabbo]

Ok, so perhaps we (myself and my fellow BFL detractors) were wrong about BFL. It's nowhere near as powerful as the Avalon ASICs, but hey, they hit their spec, so they lived up to their part of the deal. I think. Shipping these review units is a good first step, shipping the whole first run will be the true test.

 

[somini]

tutis[/url]":20vp183c]Could this kind of little box be used to brute-force password cracking attempt? Isn't it doing the same kind of math?

Came here to say this. Imagine a cluster of this.
I guess Ars will have a reference in the upcoming password article.

 

[nosensewhatsoever]

I'm pleading ignorance here. What exactly does bitcoin mining do that creates the bitcoin currency? I know I can google a lot of it, but if anyone can break it down for me real quick I'd greatly appreciate it.

 

[sonolumi]

I wonder if this box will be useful for other things once bitcoin inevitably implodes.

 

[Vossler]

zneak[/url]":1pykone8]I'll be curious to see how much bitcoins have to be worth to make up for the electricity consumption.

It is exactly why the move from FPGA to ASIC has occurred, it is simply far more cost effective to use specialized hardware to maximize profit potential by reducing the electrical bill.

 

[DrPizza]

tutis[/url]":1vj421fk]Could this kind of little box be used to brute-force password cracking attempt? Isn't it doing the same kind of math?

If not this box, then something broadly similar in concept. Most passwords aren't hashed with SHA-256, so it wouldn't be directly useful, but equivalent boxes doing SHA-1 or MD5 sure would be.

 

[chalex]

According to my trusty Kill-A-Watt, the miner is drawing a pretty constant 50 watts ... According to MacMiner, the ASIC is generating a fair amount of heat, too...

I wonder how much heat it's generating

http://www.bitcoinx.com/profit/ is a good calculator of profitability, put in 5000 MH/s and 50W.

 

[Bboner]

zneak[/url]":33oezm2e]I'll be curious to see how much bitcoins have to be worth to make up for the electricity consumption.

Since this unit only uses 50 watts to hash at 5Ghs/sec, it is far more efficient than the GPU miners that have been working at this for years using 300-600 watts to hash at 800Mhs/sec. Mining wouldn't be profitable if bitcoins were only worth $3-$5 with high electricity costs, but with bitcoins now being traded at over $100, the BFL devices are quite efficient little moneymakers.

 

[stopher2475]

zneak[/url]":26j9f78y]I'll be curious to see how much bitcoins have to be worth to make up for the electricity consumption.

If you're mining in the winter, at least you can make use of the heat and it's not totally wasted.

 

[tipoo]

That's an impressive heatsink for just a 30W product, I read that they designed it for a higher wattage part (under that heatsink you'll find a board with lots of empty spots where the scalable chips would have gone).

Edit: Ah, Ars measured 50 watts. Huh. They were initially saying 5w or something crazy low for the 5GH/s miner, that's a huge underestimate. But 50w for 5GH/s is still obviously far beyond non dedicated hardware.

 

[MoFoQ]

DerHabbo[/url]":25rbipe6]Ok, so perhaps we (myself and my fellow BFL detractors) were wrong about BFL. It's nowhere near as powerful as the Avalon ASICs, but hey, they hit their spec, so they lived up to their part of the deal. I think. Shipping these review units is a good first step, shipping the whole first run will be the true test.

yea, the 5GH/s version is the low-end unit (Jalapeno if I recall) for those who want to test the "waters" per-say.

the bigger units are where the BFL chips shines more.

 

[secretmanofagent]

Question on bitcoin: What is it actually calculating? I always wondered if it could be some foreign power's way of getting a supercomputer while making it look legit.

 

[gusgizmo]

Bitcoin is all about calculating SHA-2 (SHA-256) hashes. The applications for this are very limited, I doubt one of these would be flexible enough to be very useful as a password cracker.

 

[tipoo]

secretmanofagent[/url]":2s7l53og]Question on bitcoin: What is it actually calculating? I always wondered if it could be some foreign power's way of getting a supercomputer while making it look legit.

Really nothing, the "solution" is already known, it's almost like a huge game of battleship. You guess certain hashes, network says yes or no. It's not like breaking into anything secretly, the algorithms are publicly known and have to be for this to work.

 

[Pyros]

DrPizza[/url]":176dz220]

tutis[/url]":176dz220]Could this kind of little box be used to brute-force password cracking attempt? Isn't it doing the same kind of math?

If not this box, then something broadly similar in concept. Most passwords aren't hashed with SHA-256, so it wouldn't be directly useful, but equivalent boxes doing SHA-1 or MD5 sure would be.

This is one advantage that FPGA based solutions will always have. Different firmware for different tasks. Also, if new optimizations are made its just a firmware download away. I'm not saying its more efficient, just infinately more futureproof and useful in the long run.

 

[caustictoast]

DrPizza[/url]":mnxiylnu]

tutis[/url]":mnxiylnu]Could this kind of little box be used to brute-force password cracking attempt? Isn't it doing the same kind of math?

If not this box, then something broadly similar in concept. Most passwords aren't hashed with SHA-256, so it wouldn't be directly useful, but equivalent boxes doing SHA-1 or MD5 sure would be.

The problem is these chips were designed with SHA256 in mind. It's using an Asic so there's no way to change. This thing can mine bitcoin (or other SHA256 coins) and that's it. it'd cost millions of dollars to develop a chip to do password cracking just like the millions that went into this chip.

 

[tipoo]

I wonder though if one of these would be worth getting, as when the ASICs all start hitting the market (including massively more powerful ones), the total bitcoin mining power out there will go up extremely fast and mining coins will get much harder. Just like the switch from CPU to GPU, it became like a nuclear arms race, the little guys mining on cheap GPUs or CPUs were all rendered ineffective.

 

[rainsford]

nosensewhatsoever[/url]":17y0idzg]I'm pleading ignorance here. What exactly does bitcoin mining do that creates the bitcoin currency? I know I can google a lot of it, but if anyone can break it down for me real quick I'd greatly appreciate it.

The somewhat quick explanation is that bitcoin mining consists of trying to hash certain known pieces of data (bitcoin transactions) combined with a value you get to specify in order to produce a hash result in a certain limited range.

Because of how cryptographic hash functions work, the only way to produce hashes in that range is to try different input values (by varying the value you control, mentioned previously) until the hash happens to fall in the right range. While this can take a fair number of tries, once you know the input, you can give it to anyone else who can very quickly verify that you produced an input that gives the right output.

Because figuring out the input value takes a predictable amount of tries to do, this is known as a proof-of-work algorithm. In other words, you're proving you did the work because the only way for you to know the right input is to do the work of brute forcing it. This has a lot of applications beyond Bitcoin.

In Bitcoin, one of the values you're hashing (in addition to the value you specify) is a Bitcoin transaction giving yourself a fixed number of Bitcoins (25 at the moment). If you find a hash of this transaction, the random value you specify and the current transactions (collectively known as the block) before anyone else, you publish this information to the Bitcoin network. Each node can verify that you found a good block, and it gets added to the chain. And you get awarded the 25 bitcoin transaction you put in the block.

There's a lot more to it than that, but that's the basics. Bitcoin mining is basically cryptographically demonstrating you did work before anyone else, which gives you the right to award yourself some bitcoins.

 

[MancombSeepgood]

Where does the fan even go on that thing? I don't see nearly enough holes to create airflow through the heatsink.

 

[Swarley]

Dilbert[/url]":1z372jji]

caw[/url]":1z372jji]Is the outlet voltage in your house so bad that it was necessary to report that power AND current are constant? My condolences to all your appliances

Witty and insightful and yet downvoted. Downvoted by people who don't understand the post, yet again on Ars, so they downvote. If you don't know, skip it over, or look it up and learn. Those are your two choices.

Those are the only two? It's impossible that somebody understood his joke and thought it was dumb? I didn't downvote it, but if my computer could have seen me roll my eyes it might have done it on my behalf.

 

[IBM650]

caw[/url]"5wv2ouv]Is the outlet voltage in your house so bad that it was necessary to report that power AND current are constant? My condolences to all your appliances

Not being a mind reader, I don't know what the author meant.

BUT - The power consumed by a device is equal to the voltage times the current only if it is a purely resistive load - which this device is obviously not. For example, you could put a lossless capacitor across the power line that draws one amp of current but consumes no power (except for resistive losses in the power lines leading to it). This is not an uncommon real world experience; your old Dazor fluorescent desk lamp may be pulling over an amp of current at 115VAC while consuming only 50 watts of power. This device would have a power factor of less than 50%. How is this possible? It happens in any AC device where the current and voltage are not perfectly in phase. See:
http://en.wikipedia.org/wiki/Volt-ampere_reactive

 

[punkideas]

caustictoast[/url]":zn31gruj]

DrPizza[/url]":zn31gruj]

tutis[/url]":zn31gruj]Could this kind of little box be used to brute-force password cracking attempt? Isn't it doing the same kind of math?

If not this box, then something broadly similar in concept. Most passwords aren't hashed with SHA-256, so it wouldn't be directly useful, but equivalent boxes doing SHA-1 or MD5 sure would be.

The problem is these chips were designed with SHA256 in mind. It's using an Asic so there's no way to change. This thing can mine bitcoin (or other SHA256 coins) and that's it. it'd cost millions of dollars to develop a chip to do password cracking just like the millions that went into this chip.

The bigger issue is that even if the passwords were SHA256, AFAIK the ASICs are designed to take inputs of the beginning of the part of the block that gets hashed along with a startpoint for a chunk of the nonce to attach. It then bulk-hashes a set of possible candidates by doing a certain sequential block of nonces all at once. So in short, these are super-specific to Bitcoin or other SHA256 cryptocurrencies. I don't think it would be possible to rig them for password cracking.

 

[issor]

Wow, 26 comments and nobody has erroneously proclaimed that this will cause an influx of bitcoins (and a market collapse) due to faster miners. I'm impressed, seriously.

 

[tipoo]

MancombSeepgood[/url]":28hsm1ga]Where does the fan even go on that thing? I don't see nearly enough holes to create airflow through the heatsink.

There are vents on the top of the back and the bottom of the front. The fan faces up I believe, drawing in air from the bottom front and expelling out the top rear. The 360 slim also uses that reverse design.

https://lh6.googleusercontent.com/-aNaL ... G_7151.JPG

http://www.247btc.com/wp-content/upload ... e3_1_1.jpg

 

[gigaplex]

chalex[/url]":117z1x4v]

According to my trusty Kill-A-Watt, the miner is drawing a pretty constant 50 watts ... According to MacMiner, the ASIC is generating a fair amount of heat, too...

I wonder how much heat it's generating

http://www.bitcoinx.com/profit/ is a good calculator of profitability, put in 5000 MH/s and 50W.

My guess would be in the ballpark of 50W of heat... you know... physics and all that.

 

[GFKBill]

chalex[/url]":1kk35595]
http://www.bitcoinx.com/profit/ is a good calculator of profitability, put in 5000 MH/s and 50W.

$2130 in 3 months including cost of box? If the box is $275.

Not bad!

 

[Hinton]

zneak[/url]":1z6bth9u]I'll be curious to see how much bitcoins have to be worth to make up for the electricity consumption.

The whole point of this is that Bitcoins are vastly more worth than the electricity consumed.

Why else create something like this that does nothing else? (edit: Well, it also makes noise and generates heat, in fairness).


This won't go on forever though, as the amount of Bitcoins created are constant over time. Graphics card will forexample soon be wothless for Bitcoin mining.

 

[whisk3rs]

So if you manage to get your paws on the 50GH/s miner, even at 100W consumption, and 80$/BC, you're bound to make $16K in 3 months? What am I missing here?

http://dev.bitcoinx.com/profit/

 

[Hinton]

DerHabbo[/url]":1iu958m2]Ok, so perhaps we (myself and my fellow BFL detractors) were wrong about BFL. It's nowhere near as powerful as the Avalon ASICs, but hey, they hit their spec, so they lived up to their part of the deal. I think. Shipping these review units is a good first step, shipping the whole first run will be the true test.

[my bold]

Uh, no they didn't. The Jalapeño was supposed to be passively cooled, vastly smaller than this, and use 1/10 the power. Also it was supposed to be delivered 6 months ago.

They failed at everything they promised, except eventually delivering something. Most people still haven't received a unit though, and with 400.000 Avalon chips comming online /now/, the Butterfly labs units might not be worth much.

 

[Ostracus]

stopher2475[/url]":378mdib9]

zneak[/url]":378mdib9]I'll be curious to see how much bitcoins have to be worth to make up for the electricity consumption.

If you're mining in the winter, at least you can make use of the heat and it's not totally wasted.

Put a hot plate on top and one could cook breakfast.

 

[GFKBill]

whisk3rs[/url]":2w53zups]So if you manage to get your paws on the 50GH/s miner, even at 100W consumption, and 80$/BC, you're bound to make $16K in 3 months? What am I missing here?

If the 5GH model is 50W, I'd be willing to bet the power is more like 500W for the 50GH. But it still comes out pretty darn good, the power just isn't that big a component.

But what's with them not publishing power specs? That just seems strange to the point of suspicious.

 

[Doubletwist]

whisk3rs[/url]":3dho5aav]So if you manage to get your paws on the 50GH/s miner, even at 100W consumption, and 80$/BC, you're bound to make $16K in 3 months? What am I missing here?

http://dev.bitcoinx.com/profit/

The main thing you're missing is that as more ASIC miners come online, the difficulty will increase quickly, thus lowering your profit. Also if 5Ghashes use 50W, then perhaps the 50Ghash version will use 500W, not 100W.
They should still be profitable for awhile, depending on how many Avalon ASICs come online.

 

[whisk3rs]

Doubletwist[/url]":dke1ll05]

whisk3rs[/url]":dke1ll05]So if you manage to get your paws on the 50GH/s miner, even at 100W consumption, and 80$/BC, you're bound to make $16K in 3 months? What am I missing here?

http://dev.bitcoinx.com/profit/

The main thing you're missing is that as more ASIC miners come online, the difficulty will increase quickly, thus lowering your profit. Also if 5Ghashes use 50W, then perhaps the 50Ghash version will use 500W, not 100W.
They should still be profitable for awhile, depending on how many Avalon ASICs come online.

Got it, thanks. So, in a chase for a quick buck, if one were to preorder now, by the time the box gets delivered, it may not be super-profitable anymore.

 

[Jeff3F]

Also, to mitigate risk one has to cash out quickly, or risk inflation or problems caused by regulation, or something else like digital relatives coming out of the digital woodwork requesting facebook gifts or golden pokes or something...

 

[cpengr]

DerHabbo[/url]":kuri2687]It's nowhere near as powerful as the Avalon ASICs

This is their least powerful unit (formerly called "Jalapeno"). See their products page.

but hey, they hit their spec, so they lived up to their part of the deal. I think. Shipping these review units is a good first step, shipping the whole first run will be the true test.

The info is pretty scattered, so this is mostly from memory. I welcome any correction.

They have committed to delivering the hash rates that people ordered, but they're way over the expected power budget. So much so that they had to redesign the boards, and increase the case sizes to accommodate. The 5GH/s & 25/30 GH/s miners are going in the cases that the 25/30 & 50/60 GH/s units were supposed to go in. The cases for the 50/60 GH/s miners look to have been about doubled in size.

I write 25/30 and 50/60 above because the specs changed however many weeks ago when they roughly doubled the prices. They've still committed to delivering on the promised hashrates IIRC (EDIT: source), such that those who preordered before the spec change will get the higher rates.

I don't know if they had actually committed to the efficiency figures (GH/s / W) as a spec, but they did recently have those who preordered confirm that they still wanted to receive the product. From the email:

The key issue has been the engineering related to accommodating larger power draws than expected. A good example is the Jalapeno product. It was originally designed to be powered by USB but now consumes the power of a small light bulb (30w).

Consequently the power regulator, enclosure, airflow and PCB needed upgrading to suit. Although we are *very* aware of the undesirable dynamics of any delay, we were nonetheless obligated to make these updates in order to deliver a reliable product at the expected performance. The same adjustments have been made with all products in the lineup. You can see the adjusted product cases in our currently posted product lineup. (The Mini Rig case will be double shipped to satisfy their orders which is why we've run out Mini Rig enclosure stock).

http://bflupdate.info/2013/01/shipping- ... from-josh/