Topic: Mini Rig announcement by Butterfly Labs - 25gh/s - page 9. (Read 34235 times)

True, true.  I'm just spoiled.

The only thing required to run cgminer/ufasoft is a linux kernel and some third-partie libraries installed.
Since they are console applications, no special UI is needed. One can access it through SSH remotely.

Should anyone be interested in BAMT or other linux distros, then they must be recompiled for ARM architecture.
We haven't look into any of that yet, but we will at certain point in the future...


Good Luck,

I think it's time somebody gave BFL a pat on the back. It's impressive that they put together a complete package for $600. Custom processors, custom enclosure, custom heatsink... that's a lot of engineering work. It sounds like everybody else is gonna have a very hard time matching the new BFL processors.

For how long?

So will a linuxcoin or bamt run on a ras pi(or does it need to be ported for arm arch.) and then we can run the mini rig off that?  Is BFL going to release their own os for this including easy miner? This sounds awesome but i am curious.

My electricity is $0.00 and I have no VAT. Let's work out a rev share. I'm happy to host. I'd have immediate openings for ~Mini Rig Boxes

The individual boards are fitted with functional USB & 5.5mm barrel Power jack.  However, the in box connectivity & power are provided by a proprietary link system connecting to the Motherboard.  If you remove a card, you can run it on it's own like a Single.  Singles can also be run inside the Mini Rig enclosure, but only via USB.  

This hasn't been mentioned yet but the Mini Rig comes with a Raspberry Pi mounting rack for self hosting capability.  This would also be used for running non Mini-Rig link enabled PCB's.  

Regarding the airflow, I appreciate your thoughts.  I think it's a good idea.  We came to a somewhat similar conclusion.  The arrangement illustration isn't an accurate representation of the final design, it's just an an illustration of the modular design.  The key difference is inverted trays with baffles to trap 2 levels of hot air in 2 powered air exhaust  channels.  

Resale value is the biggest unknown with dedicated hardware. A GPU will always have resale value, even an FPGA development board adapted as miner certainly will have some resale value. However the resale value of a dedicated FPGA bitcoin miner depends completely on the value of bitcoin. So it's most straightforward to consider no resale value for a BFL miner: if the bitcoin value is high enough I won't sell my BFL miner but use it for making profit; if I don't make profit with it anymore, it's just luck if anyone will buy it (at a meaningful price), I'd be better off looking for another crypo-currency application to use it for.

What you really should be considering is resale value and downtime. If you factor resale cost in your your M/Hash cost your cost will be around .50 per M/Hash if you sell within the first two years. It's almost what BFL is selling their rig at. The only difference is that you your money will be tied up for almost 4 months based on BFL's performance to date. That's a loss of $8,000 in productivity for downtime on currently exchange rates and difficulty. You are paying 10 cents per M/Hash and getting a significantly lower power cost.  You also are investing the entire 15k betting that the market will be favorable.  If you pay a lot for electricity and don't care about the risk in BTC then the 15k won't seem unreasonable. If you have low electricity costs and want to mitigate your risk by being able to resale the GPUs if you shutdown your rig then it's not a good choice for you.

I personally am investing only 20% of my budget in FPGA hardware. I'm doing that on the assumption that FPGA will eventually be significantly lower cost then GPUs but we are not quite there yet.

Hope that helps...

it's not called business venture no matter which way you look at it... it's a hardware purchase. BFL isn't selling you a franchise. If you run a business from the hardware you buy from them it's your problem but that isn't what they deal in or what they are advertising.

There's no credit card option for a mini rig. Also, this type of transaction would be considered a business venture and it would be unlikely you'd win your charge back, especially after 15 weeks (6 months?). Same goes for the singles. If you started a 600 charge back within 60 days you'd be ok. But if you spent 15k, you'd have a lot of trouble and would probably lose. Recovering bank wire funds after 15+ weeks isn't going to happen.

This is true. Any other industry the company would have to assume some of the risk to be able to compete in price. Once again though, its a matter of needing a competitor thats hungry.

and if bitcoin went south today every single person without exception(except those that paid in BTC) would pull their funds from them by charging back their cards since unless you shipped(with tracking) the item the transaction charge back is very easy to win...

Don't know which country you're from(I know I am just being sarcastic) but here in america the only business that pushes all the risk of all their crap ventures to the consumer without any arguing and whom doesn't have to deal with any of the consequences is the the government since they print all the money and have most of the guns.

I would think that BFL knows bitcoin mining is a rather risky venture. If they from $200k to build a bunch of mini rigs and the network goes south, all they have is deposits. They are putting the risk in the right place.

This is a sore subject.  BFL's response to partial payment until delivery is that they use the entire funds to support production.  Whether this is true...who knows.

Seems they are very risk averse to fronting capital to run their operation.

and with the dust filters it becomes an air purifier :p And like D&T said a reasonable deposit would be much nicer than coughing up and locking up 15 grand for 3-6+ months, ~2K down and remainder when ready to ship is totally fair deal, BFL loses nothing by doing it since they don't ship until getting the rest of it and you don't get all that money locked away on someone else's bank account. Even at lousy 0.74% 6 month CD is a much better deal than to lock 15K at 0% interest for around that much time or longer depending on if it they actually get the Mini out in the 3-4 months like advertised.

I hope it has dust filters, high amount of airflow will turn it into a mini vacuum cleaner.

Without access to the boards and some testing it is hard to say for sure.  Maybe their design is optimal but it would seem to be.  Even non-optimal maybe the design has enough airflow to keep everything coo.

If I purchase a mini-rig the first thing I would do it open it up.  I actually had little interest in the rig or mini-rig until finding out it looks like the boards are independent.  Hopefully BFL can confirm that.  BFL can the boards function removed from the rig box and connected to a host PC via USB just like the single's do?  If they boards can operate independently and they allowed pre-ordering with a deposit (say $2K) I would order today but I don't like locking up $15K on a product which hasn't even been built yet and some questions remain.

The first though I had when looking at the CAD is the 18 individual fans are counter productive to the airflow going across the boards.  I would experiment by turning all the board on their sides and using pusher and puller fans to try and force airflow front to back.  Maybe my time in datacenters is causing tunnel vision but that type of design allows multicore database servers to remain cool under higher load.  Cold air in the front, hot air out the back.

The heatsink chosen is non-optimal for that kind of airflow but it still may work depending on heat load.  Preferably one would want a heat sink with has large number of fins aligned with the direction of airflow so cooling is achieved by pushing airflow across the finds.

Something like (top view)

B = one board placed perpendicular.  View is top view so you are looking at the sides of the boards.
B

3 ranks of 6 boards = 18 boards (although no room for expansion).
You would want a baffle or shroud to force the air past the heatsinks.  So the exact layout would require some testing, experimentation.


I would imagine you could fit it in a standard 19" rack chassis.  Maybe 4U?.  I wouldn't put 10 in a datacenter rack but maybe 8 with 1U between servers.  200GH/s per rack?  

Anyways that was my first thought.

No, not really, and it wouldnt serve a purpose, unless you mean a new "shared cooler" that uses one large radiator and links all the individual chips with heatpipes. Probably a bit cumbersome and expensive, and pretty much out of the question for DIY. I mean, it can be done, if you are a good plumber, but.. I wouldnt . Watercooling is probably easier and more efficient.

 What would be useful, but depends on the design of the heatsink, is creating a shroud that covers the heatsinks and forces air through all of them. Thats how most high density servers are cooled, there is a row of fans at the front and the air is guided over and through the cpu heatsinks with a simple plastic shroud. Not all that different from gpus actually.

So just remove the individual fans, beef up the back and front fans and force the air through the individual heatsinks. That does require the fins are properly aligned to allow that.

rjk, thanks for your input. I know BFL is eager to get feedback from the community and I appreciate you taking the time to write this up. Hopefully the current heatpipe for the rev3 can work to remove enough heat from the super single chips.

We have already seen evidence that BFL is listening to the community (mini rigs instead of rig boxes, off-the-shelf PSUs to power them) so the more constructive feedback we can get to them, the faster they can assimilate these ideas into the products.

I do have a question though. Can the current heat pipes be linked together to better dissipate heat? I'm by no means a technical expert in this area, just wondering if there is a way to use the heatpipes currently produced to better achieve "larger heatsinks".