LMGTFY


May 07, 2011, 08:50:04 AM 

I appreciate you've possibly found a buyer for all 10 slots, however I'm registering my interest anyway :) I'd be looking at between 1 and 3 slots @ 30 BTC (i.e. I'd pay you 30  90 BTC).
You could maybe consider getting a beefier machine, and selling 20 slots? Might be more cost effective than buying a second, identicallyspecced, machine. And it might suit your needs better? Anyway, consider my interest registered!

This space intentionally left blank.






Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.



joshuaMachine
Jr. Member
Offline
Activity: 42


May 07, 2011, 09:02:31 AM 

I appreciate you've possibly found a buyer for all 10 slots, however I'm registering my interest anyway :) I'd be looking at between 1 and 3 slots @ 30 BTC (i.e. I'd pay you 30  90 BTC).
You could maybe consider getting a beefier machine, and selling 20 slots? Might be more cost effective than buying a second, identicallyspecced, machine. And it might suit your needs better? Anyway, consider my interest registered!
I'm going to think about this over the weekend. If there is obvious demand for more slots, I will try to accommodate more, to a limit. There is some amount of Mh/s that I just wont be able to cross because of space and heat and power draw (it's free, but I can't break the circuit breaker). I will consider getting a machine with more PCIe slots. I would appreciate advice on a motherboard and case that could support 58+ double size PCIe 2.0 x8 or greater slots.

1Bt2G9x6fck9pAqVnDdPDrvJHUa7H4L6zE



LMGTFY


May 07, 2011, 09:17:14 AM 

I'm going to think about this over the weekend. If there is obvious demand for more slots, I will try to accommodate more, to a limit. There is some amount of Mh/s that I just wont be able to cross because of space and heat and power draw (it's free, but I can't break the circuit breaker).
I will consider getting a machine with more PCIe slots. I would appreciate advice on a motherboard and case that could support 58+ double size PCIe 2.0 x8 or greater slots.
This forum's Mrb has done some work with multiGPU systems  check out their blog, as they detail the motherboard etc they use. You'd probably need to look at 6990s rather than 5970s, but with four 6990s you could hit 2800 Mhash/s  at 100 Mhash/s per slot that's 28 slots. Might annoy your circuitbreaker, however... and I suspect that heat would start to become an issue. You'd need a mobo with four doublesized PCIe slots, but x8 isn't necessary  with a bit of hacking you could get by with x1. Mrb used PCIe extenders, but mentions that literally hacking the mobo is an alternative option (though one I'd avoid, since it's an expensive mistake if you do it wrong...)

This space intentionally left blank.



wareen


May 07, 2011, 09:48:13 AM 

90 BTC return describes a scenario where the difficulty increases to 400% of today's level TODAY. It's more likely that today the difficulty level will be approximately 100% of today's level. Do the calculation to determine the expect return at the current rate change, at a 2x rate change, and a 1/2x rate change. Report back. You will get paid the expected value of mining BTC at 100 Mh/s. The private pool choice / soloing is something you don't have to worry about. The only thing that is relevant to your BTC income is the difficulty rating.
Yes, today's difficulty level will be 100% of today's level  I agree on that But in 23 days it will be about 138% of today's level, in two weeks it will be at about 190%... My point is, that no calculation of expected return should be based on a steady difficulty level. Extrapolating the current difficulty increase rate of about 30% every ~11 days, works out to about 40BTC over a year for 100Mh/s  that is, if you had started at the beginning of the current difficulty level. If you start from the next level you won't get much above the 30BTC. Of course, nobody can predict how the difficulty will change during the next year and the current rate of increase is probably unusually high but it is a bit of a gamble. I don't want to discourage you or anybody from your plans, all I'm saying is to realize you are more or less betting on a decrease of the current miningpower growth level.




joshuaMachine
Jr. Member
Offline
Activity: 42


May 07, 2011, 10:13:26 AM 

90 BTC return describes a scenario where the difficulty increases to 400% of today's level TODAY. It's more likely that today the difficulty level will be approximately 100% of today's level. Do the calculation to determine the expect return at the current rate change, at a 2x rate change, and a 1/2x rate change. Report back. You will get paid the expected value of mining BTC at 100 Mh/s. The private pool choice / soloing is something you don't have to worry about. The only thing that is relevant to your BTC income is the difficulty rating.
Yes, today's difficulty level will be 100% of today's level  I agree on that But in 23 days it will be about 138% of today's level, in two weeks it will be at about 190%... My point is, that no calculation of expected return should be based on a steady difficulty level. Extrapolating the current difficulty increase rate of about 30% every ~11 days, works out to about 40BTC over a year for 100Mh/s  that is, if you had started at the beginning of the current difficulty level. If you start from the next level you won't get much above the 30BTC. Of course, nobody can predict how the difficulty will change during the next year and the current rate of increase is probably unusually high but it is a bit of a gamble. I don't want to discourage you or anybody from your plans, all I'm saying is to realize you are more or less betting on a decrease of the current miningpower growth level. I completely agree. Based on your calculations (I did not double check them)  if you think the difficulty level will increase by 30% every 11 days for the next year (which is already an unusually fast recent growth rate), you will STILL make 30% on your money, which is 3 times as good as the stock market long term average. I think pretty much everyone figured out that this investment's value is directly tied to the value of the Mh/s, which was made about as explicit as possible. It's in the thread title . The only reason I mentioned the 25% number is that I heard a projection that by 6 months from now we would be at 25%. Even if you only got in 6 months of mining at 25% capacity (starting now) you would end up with 45 BTC at the end of 6 months. Will we have 25% capacity in 6 months? more? less? there is some risk involved in making that guess, that's the entire point. I talked about it at length.

1Bt2G9x6fck9pAqVnDdPDrvJHUa7H4L6zE



wareen


May 07, 2011, 11:01:14 AM 

I completely agree. Based on your calculations (I did not double check them)  if you think the difficulty level will increase by 30% every 11 days for the next year (which is already an unusually fast recent growth rate), you will STILL make 30% on your money, which is 3 times as good as the stock market long term average. Yes, 30% gain if you had started a week ago, about 0% if you start in 2 days and ca. 21% if you start in two weeks. Even if you think this growth will slow down it is not very reasonable to just use some fixed difficulty level for your calculations. Anyway, the point I was trying to get across is, that time is an issue here: no matter _how_ fast difficulty increases, the time in the beginning will be the most profitable (except you count on it decreasing again). Which brings me back to my original question: when do you think you could have your machine up and mining? The only reason I mentioned the 25% number is that I heard a projection that by 6 months from now we would be at 25%. Even if you only got in 6 months of mining at 25% capacity (starting now) you would end up with 45 BTC at the end of 6 months. Will we have 25% capacity in 6 months? more? less? there is some risk involved in making that guess, that's the entire point. I talked about it at length.
25% mining efficiency in 6 months is pretty optimistic IMHO  that would boil down to an average of only 0.77% network growth per day, whereas we are more at 2.5% right now. But maybe I'm just too pessimistic Anyway, it would be great if you could estimate how fast you think you can get started.




joshuaMachine
Jr. Member
Offline
Activity: 42


May 07, 2011, 11:17:39 AM 

I'm starting the first 10 shares approximately Wednesday. I agree that with this, as with anything, you will have to make a prediction about your expectation before making the investment. If you think the growth rate will stay large or increase more, it makes no sense. If you think the large growth rate as of recent is a peak, it might. It's entirely up to you. That is your assessment to make, and no one can predict the future with certainty. Make your best guess. Weigh the risk vs. the reward. Use your brain. All those good things. Obviously, I'm not trying to make that prediction for anyone. Go run some numbers and see if it makes sense, use a variety of growth changes and see how the money works out.
I'm not sure how soon I will be able to add additional shares or at what price. The absolute earliest would be Wednesday.

1Bt2G9x6fck9pAqVnDdPDrvJHUa7H4L6zE



nster


May 07, 2011, 11:53:39 AM 

US Wall Sockets are pitiful Check your circuit breaker thing.... You better know, as some are 10A for ea socket, @ 120V = 1200W * 80% effenciency = 960W so to be safe maybe 920W. That's minimum obviously. Most are 15A 120V per socket. thats 1800W * 80% = 1440W, and usually you want to be 1 AMP below at least so 1300~1350W. You could have 20A 125V sockets, but I doubt it, which offers 2500W * 80% = 2000W so maybe 1900W If you think you have standard sockets, DO NOT go over 1500W DC. 2* 750W PSUs are usually more effencient then 1* 750W. An XFX 750W BE or Corsair HX750W/AX750W (ie: 80 Plus Silver or Gold) type quality is essential if you want to make sure your socket is safe. With 4* OCed 6990s (OCed only to 830Mhz and lowering mem clock), you can get 25 slots (the 3 overhead should go in your pocket and are just in case scenarios) can be run by 2* 750W PSUs on the same socket OR just put them in different sockets and don't expand too much. If you have that 1KW ordered, NP, get any PSU that could run 2* 6990s alone (650W~750W), 80 plus obviously, and run the 1KW with the whole system + 2 6990s ad the 750W with 2 6990s

167q1CHgVjzLCwQwQvJ3tRMUCrjfqvSznd Donations are welcome Please be kind if I helped



