I knew you were nuts, but I didn't know you were that crazy! BTW, I don't mine just for profit, it's also a hobby. If I didn't like tweaking these things I wouldn't have bought them in the first place. I just would've kept mining happily and easily with my Ant S3s.

Just take a look: (CLICK) Do you think BCP19 is a native Finnish speaker/writer?

...

So here's the science part for you guys. I only ran short tests because I know by now how the stats even out in the long run.

There's no difference whatsoever in the heat and power consumption between Linux/cgminer 4.6.1 and Windows 8.1/64bit and the custom BFGminer 4.2.0 when I swap the usb cables between the two PCs. I suspect there would be a difference in the power draw if I power cycled the units first. It would seem that if I boot them up when cold and with a powerful PSU the 700 GH unit gets too excited and overheats. When I power cycle it when warm, the power consumption and hashrate aren't so big. The 700 Gh unit takes 525 Watts right now and that's fine for me. When I booted it up for the first time after I got it from the post office it drew 570W from the wall.

Two Monarchs on Windows:



550 GH on Linux/cg4.6.1:



700 GH on Linux/cg4.6.1:



So the bottom line is that cgminer indicates way way more hashrate compared to bfgminer with the same power consumption. I don't get why the avg short time / long time hashrate stats seem to be backwards between the two monarchs on BFGminer.

I would have to run a long test to see what difference the pools see in the hashrate between cg/bfg. Could cgminer really be doing way more work with the same power consumption?