Seems like ASIC miners are about 50x more efficient than FPGA miners in both $/hashrate and $/watt.
50 * the current network hashrate of 24 terahashs/s is 1.2 petahashs/s. So when will we get there?
GPU mining is 200+ times faster than CPU mining.
During mass CPU mining the difficulty was, what, 50k-100k? Therefore the current network difficulty must be around 10-20 million since we switched to GPUs/FPGAs, right?
Analogy aside, just because we (as a mining collective) are switching to a faster, more efficient mining device
does not mean we will see an equal jump in network difficulty. I'm not saying we
won't ever see 1PH/s, but it surely won't be within the next 12 months. Even if you take all current (known) preorders from BFL and Avalon, double it all, and then add in 100TH/s for ASICMiner, you'll still only come in around 350TH/s.
There are a number of factors here, but the largest is the barrier to entry for the "Average Joe". For an average person, using some opensource software to utilize their CPU/GPU for mining was easy. With the switch to ASIC, we're now pushing people to purchase specific devices (that they otherwise might not have bought) if they want to keep mining
while maintaining profitability [obviously you can still mine with a CPU/GPU, but 300MH/s on a 350TH/s network isn't going to do much].
So something has to give..either ASIC devices become cheap as dirty, or the BTC/USD is so high that no one can resist the allure of "making a quick buck".
Not sure where you get your numbers from. Taking a quick look at historic charts, difficulty was
1 until about January '10 (1st year), it climbed to about
20 by July '10, which I believe is when GPU mining was introduced (someone correct me if I'm wrong), when it promptly skyrocketed. Reached 50,000 only in about Feb '11, when GPU mining had been around for ~6months, and was beginning to become more widely popularized on the internet.
So all in all, ~3mil / 20 = 150,000x increase. But I agree, the analogy between CPU and GPU fails, for different reasons. CPU -> GPU mining coincided also with an increase in bitcoin popularity unrelated to the technology (people had CPUs and GPUs, either would have worked for technophiles). Although there was an increase in efficiency (Hash/W) there was also a raw power factor that was introduced with GPUs as well, the opposite of which is happening with ASICs. Proven, readily available tech vs. bleeding edge DIY (kinda) startup tech, etc.
However you look at it though, in my opinion, the time to 1PH will be determined by 2 factors
1) How fast ASIC mfgs can actually get the damn things out the door.
2) Price per coin.
There was a huge huge huge goldrush to bitcoin when it was being first discovered, at low diff low value, without exchanges, or bitcoin economy. To think that people will sat back on their haunches with money to be made now is only wishful thinking to me.
The bitcoin economy is roughly estimated at 300M USD today, 1PH requires an investment of approx 21M USD. If manufacturers could supply it, and no one else did it, I would myself
