Let's recap on what we've seen in the past few months

[Bitcoin Swami]

Would a fair estimate be 200 active FPGA mining boards exist today and that would be about 20 gh/s?
If so they do represent a tiny fraction of the mining power today and their uptake would have to
increase in such a dramatic fashion to even become a blip on a chart.

Don't they do around 100 mh/s?

Wouldn't that be just 2 gh?

[1713867484]

1713867484

[Swishercutter]

Would a fair estimate be 200 active FPGA mining boards exist today and that would be about 20 gh/s?
If so they do represent a tiny fraction of the mining power today and their uptake would have to
increase in such a dramatic fashion to even become a blip on a chart.

Don't they do around 100 mh/s?

Wouldn't that be just 2 gh?

100,000,000 x 200= 20,000,000,000 

[Synaptic]

Would a fair estimate be 200 active FPGA mining boards exist today and that would be about 20 gh/s?
If so they do represent a tiny fraction of the mining power today and their uptake would have to
increase in such a dramatic fashion to even become a blip on a chart.

Don't they do around 100 mh/s?

Wouldn't that be just 2 gh?

100,000,000 x 200= 20,000,000,000 

Uh, buddy...kinda went a whole fucking of order of magnitude overboard there...

But, this community shouldn't surprise anyone by now...

[grod]

defxor more than gets it.

Those who do not, perform this thought experiment.  Assume the price of bitcoin does NOT increase, but in fact decreases to around $1.  Now think about what difficulty will do with the current GPU price to produce a coin of around $3.

You will see that while FPGA mining makes little sense at the current price and difficulty it is inevitable if price falls low enough.  There plenty of reason to expect diffiuculty to follow price, so don't worry about relative performance vs GPU.  As the profit from GPU mining goes negative mining hardware will have to become more efficient.  In spite of lower performance it will be the only practical way to obtain a piece of 7200 bitcoins a day.  It doesn't matter if GPUs will produce 10x or 100x coins at a 50% loss per coin -- that will only be practical for the "free power" miners, and "free power" doesn't scale.

ROI is still TBD.  The current generation of dedicated mining tech is not price competitive -- but if you look closely you can see promise of pricing equivalent to current GPUs.  The first outfit to hit a $100 for 100 mhash price point at ~10 watts will make for millions in sales.  77 boards sold at $610 for < 200 mhash at 15 watts.  How many do you think would have sold at $300 for 200 mhash?   How about $200 for 200 mhash?  How about $100 for 200 mhash?  Remember that there's no longer a requirement for an expensive gaming board or PCIe extenders or industrial sized power supplies, other than a cheap recycled laptop it's all about the dedicated hardware.

There's a good reason we all expect price deflation for technology.   Saying it'll never happen is placing just such a bet.

[Synaptic]

defxor more than gets it.

Those who do not, perform this thought experiment.  Assume the price of bitcoin does NOT increase, but in fact decreases to around $1.  Now think about what difficulty will do with the current GPU price to produce a coin of around $3.

You will see that while FPGA mining makes little sense at the current price and difficulty it is inevitable if price falls low enough.  There plenty of reason to expect diffiuculty to follow price, so don't worry about relative performance vs GPU.  As the profit from GPU mining goes negative mining hardware will have to become more efficient.  In spite of lower performance it will be the only practical way to obtain a piece of 7200 bitcoins a day.  It doesn't matter if GPUs will produce 10x or 100x coins at a 50% loss per coin -- that will only be practical for the "free power" miners, and "free power" doesn't scale.

ROI is still TBD.  The current generation of dedicated mining tech is not price competitive -- but if you look closely you can see promise of pricing equivalent to current GPUs.  The first outfit to hit a $100 for 100 mhash price point at ~10 watts will make for millions in sales.  77 boards sold at $610 for < 200 mhash at 15 watts.  How many do you think would have sold at $300 for 200 mhash?   How about $200 for 200 mhash?  How about $100 for 200 mhash?  Remember that there's no longer a requirement for an expensive gaming board or PCIe extenders or industrial sized power supplies, other than a cheap recycled laptop it's all about the dedicated hardware.

There's a good reason we all expect price deflation for technology.   Saying it'll never happen is placing just such a bet.

No, you completely forget the difficulty retarget, just like that other fool did.

Yeah, a shitload of miners will bow out when the USA price inevitably plunges further, but then difficulty will decrease proportional to meet the power of the remaining GPUs.

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...

[geek-trader]

This user is currently ignored.

I *LOVE* the ignore feature! LOVE! 

[Swishercutter]

Would a fair estimate be 200 active FPGA mining boards exist today and that would be about 20 gh/s?
If so they do represent a tiny fraction of the mining power today and their uptake would have to
increase in such a dramatic fashion to even become a blip on a chart.

Don't they do around 100 mh/s?

Wouldn't that be just 2 gh?

100,000,000 x 200= 20,000,000,000  

Uh, buddy...kinda went a whole fucking of order of magnitude overboard there...

But, this community shouldn't surprise anyone by now...

20gh/s is 20 billion hash/s.  100mh/s would be 100 million hash/s...multiply that by 200=20,000,000,000.  I don't see how that is wrong.  

[grod]

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.

[Synaptic]

Would a fair estimate be 200 active FPGA mining boards exist today and that would be about 20 gh/s?
If so they do represent a tiny fraction of the mining power today and their uptake would have to
increase in such a dramatic fashion to even become a blip on a chart.

Don't they do around 100 mh/s?

Wouldn't that be just 2 gh?

100,000,000 x 200= 20,000,000,000  

Uh, buddy...kinda went a whole fucking of order of magnitude overboard there...

But, this community shouldn't surprise anyone by now...

20gh/s is 20 billion hash/s.  100mh/s would be 100 million hash/s...multiply that by 200=20,000,000,000.  I don't see how that is wrong.  

Right, I'm sorry.

I got the hash levels mixed up, it's been quite awhile now since I've been mining.

[Synaptic]

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.

I'm following perfecly, you're just for some reason completely negating the outrageous cost per mh/s of an FPGA vs. a GPU. Yes, include electricity, but don' just ignore the unit cost of the hardware too.

You really need to check your calculations.

[fcmatt]

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?

[Synaptic]

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?

BLEEP BLEEP mofo,

and thanks for that, these people just don't get it, your post is a much milder take on my same reasoning.

But, fuck mild.

[grod]

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.

[Swishercutter]

Would a fair estimate be 200 active FPGA mining boards exist today and that would be about 20 gh/s?
If so they do represent a tiny fraction of the mining power today and their uptake would have to
increase in such a dramatic fashion to even become a blip on a chart.

Don't they do around 100 mh/s?

Wouldn't that be just 2 gh?

100,000,000 x 200= 20,000,000,000  

Uh, buddy...kinda went a whole fucking of order of magnitude overboard there...

But, this community shouldn't surprise anyone by now...

20gh/s is 20 billion hash/s.  100mh/s would be 100 million hash/s...multiply that by 200=20,000,000,000.  I don't see how that is wrong.  

Right, I'm sorry.

I got the hash levels mixed up, it's been quite awhile now since I've been mining.

Just the metric system...at least one Mars lander was ruined by not understanding it.

[fcmatt]

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.

[grod]

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.

Good point, that's a design flaw in the current iteration.  The cards would need a redesign to use their own bus so that multiple cards could be daisy chained via a single USB connection.  Or use other ways to improve the density of FPGAs per USB port. 

Then again, would be cheaper to go mini-itx with a large number of USB ports.  A $150 27 watt device with 6 usb ports to power 600 boards (read: over 100 ghash, $360k farm at retail 1 unit pricing) is probably cheaper than any other custom solution.  The routers should suffice for up to 20 ghash ($60k farm at retail 1 unit pricing).

The gating factor per power supply would be total wattage, not the number of molex connectors.  Each typical consumer $50 500 watt PSUs should be able to power 50 watt FPGAs assuming total draw of 7.5 watts per socket, some cross-loading on 5v and 3.3 and the majority of draw on the 12v.  Let's call it another $2/board for power supplies  -- so another $1400-2000 in PSUs for the $360k 100 ghash farm.

Either way, the marginal cost over and above the FPGAs themselves is still noise.  Even if I suspect the actual boards would be 1/2 to 1/3 the current retail pricing in volume and you're looking at $2000 in ancillary paraphernalia for a $100k 100 ghash installation.

[defxor]

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.

Exactly. That's why I don't get Synaptic's "price must go up for ASIC mining to take off". It's the other way around.

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

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.

this

I think it's the other way around actually. I won't even touch mining until we're at ASICs, since I consider the price to follow production cost until we hit the uptake vs deflation knee of the curve. GPU miners need to charge $6/BTC. FPGA miners can charge $0.5. ASIC miners can ...

I consider mining bitcoins to be just as mining any other form of natural resource. The price floor is at extraction cost (plus minimal profit), anything above that is speculation. Since I agree with you that speculation won't continue to drive BTC price in the near term my projection is that the floor is where the cheapest mining can be had. Currently that's FPGAs, but they will be replaced by ASICs. GPU miners, like CPU miners already have, will need to drop out.

Do note that there's a qualifier, the above is only true if speculation doesn't take over. We've already seen that happen once and it could very well happen again if there's suddenly a huge uptake which needs a capital influx the current 7200 BTC/day inflation can't satisfy.

[Synaptic]

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.

Exactly. That's why I don't get Synaptic's "price must go up for ASIC mining to take off". It's the other way around.

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

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.

this

I think it's the other way around actually. I won't even touch mining until we're at ASICs, since I consider the price to follow production cost until we hit the uptake vs deflation knee of the curve. GPU miners need to charge $6/BTC. FPGA miners can charge $0.5. ASIC miners can ...

I consider mining bitcoins to be just as mining any other form of natural resource. The price floor is at extraction cost (plus minimal profit), anything above that is speculation. Since I agree with you that speculation won't continue to drive BTC price in the near term my projection is that the floor is where the cheapest mining can be had. Currently that's FPGAs, but they will be replaced by ASICs. GPU miners, like CPU miners already have, will need to drop out.

Do note that there's a qualifier, the above is only true if speculation doesn't take over. We've already seen that happen once and it could very well happen again if there's suddenly a huge uptake which needs a capital influx the current 7200 BTC/day inflation can't satisfy.

it costs at least a million USD to develop and fab an asic in any cost effective economy of scale...

if not a million than some other large number that will never be spent on such an idiotic venture.

[Revalin]

Eventually, Bitcoin will go back up. Its fundamentals are too strong not to.

The question is when.

What fundamentals?  I consider the fundamental price to be the amount of value that needs to be stored for a few days in the (buy @ exchange > sit in wallet a few days > buy goods > merchant sells @ exchange) cycle.  That will raise the price considerably if BC becomes much more widely adopted.

But I don't think we're there yet.  The vast majority of the current price is speculation.

I believe some e-currency will eventually become extremely popular, but it might not be BitCoin.  The deflationary nature makes it very prone to fluctuating value due to speculation, and that (along with the "damned early adopters" effect) might be enough to drive the mass market to another currency (perhaps another cryptocurrency which hasn't been invented yet) which they think will be a safer store of value.

[defxor]

it costs at least a million USD to develop and fab an asic in any cost effective economy of scale...

A product like Xilinx EasyPath ($75000 NRE + lots of orders) would bring current FPGA upfront costs down quite a lot. What's needed for someone to gamble on such a venture is the belief that ASIC (or if you want to call EasyPath a hybrid FPGA/ASIC solution) miners would grab a significant share of those 7200 BTC per day. ($2/BTC is still $5.26M/year)

If the BTC price is going down to levels where it's just not profitable for GPU miners to mine due to electricity costs, that will indeed happen.

I looked at the Xilinx excel sheet yesterday. Did you?

edit: I haven't verified that $75000 is true for all FPGA revisions though, but the point stands. $5.26M/year is absolutely a market into which someone can spend hundreds of thousands if they believe they can capture a large portion of it.