Topic: Cairnsmore1 - New Short Lead Time Preorder - page 3. (Read 11660 times)

Does anyone know what size the DC plug is for the CM1? Is it 2.5 or 2.1mm?

Preorder is now closed for at least a few days whilst we sort out delivery offers.

With LTC diff increase 1G will not be enough very soon

I think that the real memory requirement would be in range of 500Mb-1Gb based on what gpus use.

There are a few possibilities and none will be as good as some of our other products with better memory interfaces. There is also a question of how much memory is actually required. I have seen a wide range of numbers from people in this forum so that is something to established.

Possibility1

CM1 contains logic only and works with a host processor that has software and memory. Bandwidth of USB will be a major limit.

Possibility2

1,2, or 3 FPGAs on CM1 are used as memory. We get about 700KB per FPGA when used as memory.

Possibility3

The up/down interface is used to support a custom memory board. There are 18 wires available to make this connection which are never going to compete with a direct DDR3 interface but the add-on board could be a very large memory size. 100GB would be possible but maybe not cheap.

Outside of these possibilites there are further hybrids of these ideas e.g. use one FPGA as a cache with larger memory as add on board.

Bear in mind the original Litecoin point was made in a different tread in the context of people thinking ASIC would kill CM1 Bitcoin mining profitability and this was something else to do with these boards it is never going to be the best solution possible. Almost any FPGA board will have Litecoin capability but not necessarily huge performance.

If you want Litecoin miner we probably already have better boards already in the product range for this purpose although prices may scare you as they never priced/designed for this market. They just happen to be possible better fits to the requirement. So far all of this conjecture. Most people have looked at what software does to do this processing and there may be entirely better ways in a FPGA than just converting a software function into FPGA hardware functions. Such radical thinking actually needs to be done by someone who knows FPGAs very well and probably software pretty well too so it is not going to be a simple or quick task to do this sort of analysis/design quickly. I am not going to promise any timelines but it is one task we are hoping to get some time on. However our design team are very loaded in the 2-3 month span and probably no way that will happen here in the short term.

Can you elaborate on that, AFAIK ltc requires some amount of ram to be mined. Cairnsmore1 does not have built in memory or any ddr3 capable slot.

We are going to temporarily close the preorder tomorrow as we are getting close to maximum numbers we can handle short term. Depending on deposit collection it may re-open in a few days. Anyone on the pre-order should get an indication of possible delivery in the next few days as we work through and confirm orders so do check you spam boxes just in case that email ends up in there.

Ok time for an update.

We are collating the pre-orders and will start to allocate boards tomorrow. We have already got some FPGAs in but the process of ensuring that we have enough FPGAs for the potential orders is still ongoing. However we can now start to contact some of you to confirm the pre-order, delivery, and take your deposits and that will start tomorrow. It is going to take us at least a few days to get through the entire list of people wanting boards and that is partially dependent on our FPGA deliveries being confirmed as well. So do have some patience whilst we work through all of this.

FRC, PPC and TRC all likely as profitable or more profitable using an FPGA... things to ponder. Litecoin aside there are alternatives to BTC. Mind you there will still be months of use in BTC if prices keep going up even if you mine at a loss and keep the coin in a few years you'd make back the cost of the unit easily.

First to say I don't think we ever promised a timeline on doing a Litecoin implementation to run on CM1 so no one should base decisions on that. We think it is possible but that is a long way from a stating an actual performance level. At the worst case memory could be accessed vis USB or up/down connections but these may not be good enough for decent Litecoin mining performance and work needs to be done on that to understand the engineering requirement. We probably won't look at Litecoin oursleves for 2-3 months based on our existing design commitments.

Anyway the purpose of this tread is inform interested parties on our progress of make CM1 available to all that want it for whatever reason not to deviate into sidelines. There are other threads for that. We are making progress on working out what we can do in the 2-6 week target timeframe and response to the large number of emails we have to deal with will come. Please be patient on this front as more effort is going into getting the answers than answering emails.

What is being questioned is not the value of "the alt-coin boat".

Although that might be questionable.

But the statements about mining 'alt-coins' with CM1 hardware.

Made right alongside statements about Bitcoins being 'incapable' of 'small' valuation usage.

-- edit

I've actually been, mostly, satisfied with my purchase of a couple of CM1's which were delivered last Fall.

But I highly disapprove of dubious / unsupported statements. (And, frankly, I expect better of "yohan".)

So, yeah, I "disapprove" of a *lot* that I read on bitcointalk.org.   

If you think BTC is the only option for an FPGA you might be missing the alt-coin boat.



Data brought to you by DUSTCOIN.COM

Unless you know of a bitstream you can flash to one of these devices *now* for Litecoin you shouldn't make a purchase decision based on assuming it *might* be able to in the future.

Just my two cents worth. (About 0.0002 BTC.)

How certain are you these will work with Litecoin?

Doesn't Litecoin mining require a signifcant amount of ram as well?
Correct me if i'm wrong.

People actually said this 9 months ago and many people didn't buy a CM1 for that reason. Anyone that did has made a nice profit. Obviously the game has changed a lot with exchange rates going up a lot and you can speculate if that rise will contine or even collapse. At the moment we are just responding to demand. If people want CM1 we will try and make them.

It is very high risk with ASIC shipment. People can take risk with top of the line equipment, this is no longer so.

Based on current difficulty, 1 month breakeven price for a Cairnsmore board is 175 USD -  not considering any further difficulty increase within this 1 month. If anyone is paying more than that, good luck recovering money in any reasonable time period.

Only good for a hobby at this price, unfortunately.

It could be.  At full power these boards only draw 45 watts.  And you can drive them with a very low watt controller.  A 7970 is going to require at least 300 watts with a system.

So these board will be cost effective a 6-7 times higher difficulty than the graphics card.  Depending on the shape of the difficulty curve between those 2 points, and the exchange rate, you could do very well.  Or painfully badly.  It's a tough call.

It was a lot harder decision at $7 bitcoin and BFL promising ASICS in 2 months.

That said, I had a buyer for 20 boards drop out on me, so I still have a few for sale for BTC only

Maybe someone can educate me: at $960 for 760-880 MH/s, wouldn't one be better off by buying 7970s?

Is this worth it because of lower power consumption?

Indeed.  For someone who failed to ship a bitcoin bitstream you are awfully confident it will work for Litecoin.

So much for professionalism from Enterpoint.