Quote from: grod on September 03, 2011, 05:34:29 AM

Quote from: fcmatt on September 03, 2011, 05:18:14 AM

Quote from: grod on September 03, 2011, 04:58:27 AM

Quote from: Synaptic on September 03, 2011, 04:23:46 AM

There is simply no point at which the difficulty will be low enough and electricity expensive enough to make FPGAs more profitable than GPUs!

Fucking christ you people need to buy more nootropic drugs and fewer bitcoins and associated paraphernalia...

Not getting through, let's try a simpler example.

If the price per BTC is $50 but cost to produce is $3 there is no real reason to optimize for cost of production. You will want to optimize for volume. If you could produce 2x as many coins at the cost to produce of $6/ea, you would do that. The actual difficulty is immaterial, what matters is the profit margin per unit.

If the cost to produce is $3 and price is $8 (as we see now) there is still little reason to focus on unit cost as the main optimization parameter.

If the cost to produce is $3 and the price is $1... Well, I've already gone into it, you're just not following along.

BLEEP BLEEP BLEEP you BLEEP BLEEP and your BLEEP math.

I am just not sure many people are willing to spend 420 dollars for a FPGA board that will make 3 dollars a month with difficulty at a million
and a BTC price of $1.

Lets day difficulty just crashes to 200,000. With a BTC price of $1 that would make 15.29 per month. 27 months until it is paid off without
accounting for electricity costs of the board and some low power PC that has a molex connector and a usb port. So we might as well
triple or quadruple the cost of electricity?

If that happens I have to wonder if bitcoin will just die?

Now you're following along as to why I think FPGA mining may be the beginning of the end.

Also, you don't need a PC. Something like a netgear 3500WR router could feed a few terahash worth of these boards. It runs linux, and has a USB port, and uses something like 5 watts. I think I paid $35 for mine. Power supplies can be stand alone, low end (inexpensive) and each molex connector could split to feed quite a few boards. The marginal cost over and above the FPGA chips themselves would be quite low.

So where I'm headed with this: a crash of difficulty because the ONLY viable way to mine is FPGA, with each board having a quite low Mhash/$ but high Mhash/joule efficiency. Further fueling a collapse in price (which makes GPUs that much more impractical for mining) until a new stability level is reached. Mining 3 coins a day at a sale price of $1/coin and power cost of $2 is not as smart as mining 1 coin at a cost of 20c.

If you look to recent investment vs price you would see people were more than willing to build $700 machines to make $5/day after power costs. That's a 3 month ROI assuming free power and a 50% depreciation of assets. I could make a guess as to where price and/or difficulty would have to be to justify the current $600 for 200 mhash at 15 watts boards, and it's certainly not $50/BTC. Nor $8/btc at 1.8M difficulty.

The big question is: how fast and how far will BTC pricing fall. If you assume a strongly bullish pricing scenario for BTC then I will concede FPGA mining will never be profitable.

This is simply nit picking but that little netgear cannot possily handle more then 100ish boards from a single usb port. Not due to
power but the spec of the usb design. So a max of 10 gh/s lets say. Or for the higher end FPGA board, double that speed. Keep in
mind that each hub counts as 1 off the max of 127 devices a single host controller can handle.

Also a single molex connector can do ?60-72 watts on average?. Lets say 6-7 of the 100 mh/s boards or 3-4 of the 200 mh/s boards.
Giving some room so we do not max it out. I am guessing here what they can do without looking.

But even with all this said, I agree it is a small fraction of what several boards would cost let alone 25 of them.

I need to chew the fat on the rest of your post before replying. It is an interesting thought exercise to figure out what happens
if the price keeps degrading and difficulty goes down with it.