There are none so blind as those who will not see....
Indeed.
Quote
The PCBs currently in hand (there were several hundred of this revision made) cannot support speeds above ~550GH with "A" chips due to heat constraints
So instead of showing 550GH to their patient customers, it's a better idea to show 350 with a 10% error rate? Right.
I'm calling BS on the 550 number, as there is no reason to publish pictures of a "Monarch hashing!" performing poorly if all it takes is 2 good chips off the shelf to show something approximating advertised performance.
By all means the particular board being shown is a test board that is not going to be sold anyway. Who cares about all the bad boards, it only takes one to demo. Why publish only pics of a bad board if there was at least one board that once did 550? How else would anyone be able to claim they can do 550 if it was never done? More "theoretical performance" ? Bullshit, bullshit and more bullshit.
Quote
Why push the board beyond it's limits just cause the chip can handle it? The increase in power to run the chips faster causes a greater current draw and therefore a higher W/GH rating. Why waste the extra electricity for a few extra GH?
The first part is funny, because the picture that was published shows a monarch that is already being run well beyond its limit. 9 percent HW errors + a bunch of rejects is evidence to that. So you were saying?
The second part is erroneous because either way with two of those mythical A chips at least you could SHOW a Monarch working at reasonable speed with reasonable error rates, without pushing the PCB any harder. Better chips run higher clocks with less voltage. Right now BFL has shown nothing but a half broken product, half a year behind schedule. True to historical form I suppose.
