I was going to switch back over to BTCGuild tonight

Hopefully you can work out whatever kinks are happening in the new server as I'd like to drop my MHash on your pool.

Is it possible to use this tool set up a backup pool? That is to say run your cards to mine for pool A, but if pool A goes down to switch to pool B until pool A goes back up?

If you are asking with regards to mining, crossfire is irrelevant for mining.

If you are just asking for gaming, then yes it will work, though iirc the 6970s will be downclocked to match the 6990 clock speeds.

On the upside I don't have to burn all my mining hardware for kindling to keep myself and my children warm at night now that I spent all our money on it and difficulty made me a poor man...

On the downside (a little more seriously) transactions are taking a lot longer now to process than before. An impatient man who is too cheap to add a transaction fee is saddened by this.

If you're referring to me, I didn't say he couldn't make a thread, I just said it was stupidly pointless to do so for such an irrelevant topic. Perhaps I should have added IMHO. As people seem to find little modifiers for some reason mollifying.

Like if I said "No offense, but this is the stupidest thread in the world", normally someone would be mad at that obvious insult, but I said no offense, making it better.

I'm glad you are happier on slushs pool, there needs to be a diversity of pools all competing their hashing power. However the PPS drop doesn't seem to be a reason to make a thread really.

I seriously don't understand why it is so difficult to comprehend. My point is that the numbers are completely baseless, hence they have no relevance of any kind. This is the opposite of what a reasoned analysis is.

Using $100,000 per bitcoin is absurd, but it is no more absurd than a difficulty that is so large it overflows the number column of the excel spreadsheet. The last difficulty his spreadsheet can even display is 7,407,009,366 on day 240. That's 7.4 Billion difficulty. What is unstated about using $100,000/bitcoin is that you are making a loss when the difficulty is 172,569,134,712.082024157046. A bitcoin rate of 100,000 dollar is unreasonable to you but a difficulty rate of 172.6 billion is reasonable? Why? Where is the reasoning to support that? Do you know how much hashing power that would represent? 434,000 difficulty represents ~4THash/sec, I don't know the calculation offhand so I could be wrong here, but if there is a 1:1 correlation then 172.6billion difficulty represents 1,590,498THash/sec. Perhaps someone will come out with super FPGA mining clusters but as it stands either bitcoin mining would have to explode into crazy realms of profitability to entice that much hashing power, or it would never ever ever happen. Perhaps the argument stems from an unspoken misunderstanding, so I will say this clearly. Difficulty is not arbitrary. It is directly related to hashing power such that it maintains roughly a bitcoin generation rate of 2016 blocks per 2 weeks.

People keep using terms like nonsense for their opposition, then they spit out nonsense. I apologize if this seems like I am merely attacking, but I am trying to get across a point here, that while return to zero profit is definitely a possibility, and something to be considered, this analysis is far from being anything factual.

for some reason this thread tickles my funny bone, think imma sub

Sure there is, for a casual miner. Pool luck evens out over long periods of time, but someone could be lucky as a casual miner by hopping on only during pools fast block breaking time, though if it were a score based pool the hash rate would have to remain constant.

Your phrasing is very confusing. You have 2 5850s, does windows(?) detect them both or only 1? If it detects them both what do you mean "unusual behavior in openCL"?

As far as mining goes you don't need the crossfire cable btw.

Of course there is the threat of hashing power becoming too concentrated. We've already seen deepbit take 50% of the total hashing power before.

Regardless, bitcoin is small potatoes right now, meaning that it can be destroyed fairly easily if someone with significant means really intended it, but also that there is not enough incentive for this to happen arbitrarily.
FPGA is not a threat right now, nor is it likely to become so for a while, there is a significant market behind GPU creation, efficiency, optimization (which bitcoin benefits from indirectly), and very very very little behind FPGA. It is possible that someone enterprising enough to engineer their own custom setup at the expense of days/weeks/months/years(?) and tens of thousands of dollars, but even then it is unlikely they will invest enough to compromise btc, where is the return?

It is certainly likely however that hashing power will become increasingly concentrated, not into the hands of a few, but as profitability diminishes the large farms will be the ones to remain (except for the hobbyists , but they will not make up a majority of the power on the network).

6870s yield terrible MHash/$. Well I shouldn't say terrible, but it is quite poor, if you buy new it's one of the few reasonable options left I suppose. Buying used is a far superior option if you can wrangle it though.

@OP the MSI 890FXA is a good board, can do 4 cards definitely though a bit pricey. If you can find a cheap board and buy some PCI extenders you can save yourself a few bucks. You don't need x16 slots, x1 slots work fine. One warning, hawks are custom coolers, they might not fit into standard 2 slot spacing.

Processor contributes essentially nothing, sempron 140s are ideal.

1000Watts is ridiculous overkill for 4x 6870s. You are essentially throwing money away unless you found a good deal on a 1000Watt PSU. One of the 6870s best features with the reduced processor count is improved power efficiency. If you downclock the memory a 750 should be plenty. Getting a high quality matching wattage PSU is far superior to getting a higher wattage junk PSU. tx750s can be found for $100 easily.

Cooling... The 4 cards are going to JAMMED together. Depending on the cooler the card is equipped with it will handle this with varying degrees of compliance. Risers or extenders are a good idea. If not, you will need to press them apart in some way and force air in between them with a fan setup of some kind. Otherwise you will have problems with sandwiched cards, crazy loud fans, overheating and so on.

I'm not sure I'd do this buy however until you calculate if the return will be worth it. Just a fair warning.

All positive hopes. Ideas which were somewhat dashed for me by this comic http://xkcd.com/893/

Bitcoin is very very very new in the larger scheme of things, and as with all fledgling ideas competing with existing systems it comes with a lot of inherent risk. If you buy a bitcoin today, tomorrow it might be totally worthless (which is possible with anything, but is less unlikely with bitcoins). I'm not sure exactly what you and your friend are arguing over, is it making money via trading, or?

You don't have to use MtGox to trade USD->BTC->USD, there are direct trades via forums, irc (#otc), in person, etc. And ideally the system when large enough would allow you to merely go USD->BTC->Goods and Services. Or even just get paid in BTC and buy stuff with them. Of course that's a ways down the road obviously if ever.

So your friend is right, who knows if bitcoin will survive the growing pains any system must go through during its infancy, lots of people seem to have faith in it, you'll have to decide for yourself. It's less of a pyramid scheme than the USD though, which just happens to be more well established.

I bet it's a hell of a lot easier to brute force the key in your head :p

Anyway, that amusement aside, I too like the concept if not the execution of this idea. Gives snoop doggs lyrics -- Got my mind on my money and my money on my mind -- fresh new meaning.

Since you want to bring semantics into it, I will use the definitions most relevant to this discussion.

Statement: a single sentence or assertion.
Argument: a process of reasoning; series of reasons.

Sorry for trying to give you the benefit of the doubt by assuming you used reason to arrive at a conclusion.

You don't think your numbers are biased? You state that bitcoin value increases to $100,000 after 360 days, yet profitability falls out after 160 days. Why would people be mining at a loss and buying coins at that price? The only absurdity is presenting numbers that are bombastic to try to make a point. It is clear from your presentation that you are leaning in a specific direction regardless of your assertion that you are just some impartial observer following the irrefutable math.

Again, if you want to use ridiculous numbers, in the past TWO MONTHS, 60 days the value of bitcoin has increased 1000%, that is the only FACT that exists (none of your projections are facts as you seem to suggest) in this analysis, and it has been completely ignored. Am I saying that will continue? Of course not, but if you are going to be pulling out numbers why are you not using what exists?

If you argue that those numbers aren't realistic, which they of course are not, then you need to add more reasoning to your analysis. And as such it has been stated multiple times in this thread already as profitability decreases so will the increasing hashrate/difficulty. Projecting static increases without taking this into account is just a way to push an agenda.


Just so I don't sound like a negative nancy:

Here is a useful tool another member created to estimate profitability, this is not in relation to buying a new rig like the OP, but it gives you a sense of where you would stand in terms of the profitability regime. It's a little more complex, and thus a little more useful.

http://forum.bitcoin.org/index.php?topic=9111.0

Really now people need to stop estimating the power use of a rig without source. 5830s are measured to use slightly more power than a 5850.

2x 5870s + 1x 5850 uses < 550 Watts from the wall all overclocked (yes I actually measured it), and remember that these PSUs are rated for power delivered to the system not drawn from the wall. 600Watts for 2x 5830s is beyond overkill, so stop saying it's not enough.

That aside, the CX600 is not a very good power supply, it can only provide 480W on the 12V rail (meaning its more like a 500W PSU). For the same money (roughly) you are far better off with the TX650 v2. Or if you really insist on being cheap the Rosewill Green 630 is a good PSU as well and cheap, and actually delivers 600W on the 12V rail.

As I understand it, hashing is not limited by PCIE in any way shape or form. There is no card that can max out even a PCI-E x1 slot.

Just make sure you have 2 spaces between the two slots, or else if you use a different card in conjunction with the mining card you won't be able to fit them. If it will be just that card in the machine then it will be fine. I will point out though that it will be very unprofitable to mine a single card on a machine.

Your thread was quite a bit more restrained than most, which makes it less entertaining but more reasonable. I will admit I personally prefer the "OMFGGGG SOMEONE BUILT A SUPER DUPER FPGA MINING CLUSTER DESTROYER AND ADDED 3THASH TO THE NETWORK WE'RE ALL DOOMED" but that's just me.

Grats to Cyde, who beat all the odds!

Keep the betting going, it was interesting while it lasted