Dude, your question shows exactly how far back you read... The whole reason this shit about bitfury cropped up is because people were claiming ridiculous numbers on the efficiency of a 65nm chip which for all intensive purposes should not have been possible and then transformed into an argument over whether or not the BF chip was more efficient than KNC's. That was the point he was getting at I believe, people where claiming some crazy shit, with out any proof to back it up...

Pencil modded to about 1.1k resistance (the actual is slightly better because the graphite reacts that way to heat), added heatsinks and fans and tuned chips speeds to lower errors below 10% and the pool stats for the past 15-20h are:


This figure includes 3440Mh/s worth of fpga hardware, so the actual hashrate of the Bitfury is 72,902 Gh/s

and the all time reject rate for my Bitfury:


;tldr; Im pushing almost 36.5Gh/s per board