Quote from: dlezama on December 04, 2017, 01:07:37 AM

Quote from: Bajula on December 04, 2017, 12:20:05 AM

Quote from: dlezama on December 03, 2017, 07:16:46 PM

Quote from: martyshmoo on December 03, 2017, 06:54:29 PM

Quote from: Marvell2 on December 03, 2017, 12:33:44 PM

Quote from: lhajnal on December 03, 2017, 10:32:32 AM

Hy guys
I don't understand, why better this firmware than the stock firmware? Because the profit on19,3Ghs with 1200w is equal 17Ghs on 750W-1,5 fee.

yeah this is the only question i have but the efficency numbers are crazy with this firmare:

stock
19600 ms at 1200 watt - .0612 watt/ms or 16.33 ms/watt

18500 ma at 920 watt - .049 watt/ms or 20 ms/Watt

15000 ma at 700 watt - .046 watt/ms or 21 mhs/watt

thats roughly a 20 percent increase in efficency which also traanlatea to longer life on the chips
which equals more time to roi , as well as less heat

now all i need are some numbers on the dev fee to see the true gains

Yes, at like 9 bucks a day, its less than a penny per d3 miner per day. and my miner makes no noise, and I can turn it down really low for the night to be low power consumption when there is no sun to collect.

still worth it.

M

30 cents a month... thats cool.

I saw this russian building an new controller alltogether for the d3 so it could mine scrypt, quark, and x13/15, etc. He wants 300 bucks a unit. I am considering this as well.

If you have 100 units I understand your concern. But a couple, this is worth it in all cases.

M

How hard is it to understand that ASICs do one and one thing only. https://en.wikipedia.org/wiki/Application-specific_integrated_circuit

I have to interject, that the specific "thing" this asic is supposed to do is 11 different algos, and yeah you aren't going to be able to add scrypt, but *IF* bitmain had a little forethought when they did the chip (which is possible since we will never see any real specs on it) they may have already done up to x15 and it just never sends the signal to use the other algos. (It is just as likely though that x11 is all it can do, but the possibility MAY be there for x13 and/or x15 ) Now because it can do all of the x11 algos .. again with the MAY it MAY be possible to trick it on the software end to mine a single algo of the 11. - Though you are 100% right about if the hardware "ain't already down with it" you ain't doin' it. So odds of adding an algo? very very small and only exist at all if the chips were already built to do other algos and the function just wasn't acessed. Odds of being able to separate out a single on of the 11 we already know are there? much greater, but again depends on alot of little factors that most of us will never know. Might be that jigging cgminer might be enough, might not be possible at all. I don't know enough about how x11 signals to change algos to the next one in the list.
So possible? maybe, but worth the effort that it would need? dunno.

Good stuff, thanks for sharing. I also thought that maybe bitmain was smart enough to make the chip a little more generic to hopefully reuse the design, mask, manufactured chips (although with the numbers they move a few million bucks may be pocket change). BUT, of course if they are smart they also put some eFuses there and the ones we have are burnt to be X11 forever Grin

Entirely possible, like lets say you design the chip with x15 kill the x13 and x15 possibilities and YOU don't have to change a thing but taking out that one little thing that kills it and voila x13, now you roll out a new pile of miners, but production costs are down no refitting machines to crank out the new chip... 'course it still could be a "hey someone is doing x11 crank em out while we can make a little $$$ out of it.. go go go!" and they never went the route of x15 and downgrading.