Topic: New firmware AsicBoost antminer S9 overclocking 17Th air cooling (Read 2447 times)

TH's means nothing!

Calculate with Diff set and shares accepted, then you can calculate right!

The shown TH's at the front is just a simple line manipulation nothing else!

https://github.com/Zwilla/bmminer/blob/dce6d9596ea83649785c9020f8b1e95680e92ea6/driver-btm-c5.c#L4529

Also offering "decrease hardware errors" is a scam!

Proof of concept?

HWE is the result of a buggy code style by Bitmain! Proof?

Rewrite the thread handling and you will see where the problem was/is!

Will publish next week my source code (BitCaine5) where Bitmain can steal again

https://bitcaine5.com

It is really easy to calculate use a pool like viabtc.com

do it  right after a diff jump is set and use pps+

you get a  guaranteed payment of  0.00004156 a day


 "Daily Yield: ≈0.00004156 BTC Per TH/s (≈PPS 97.26%)"

so 2x    s9's  without s9 firmware mods  for a day

then 2x s9's with s9 firmware mod for a day

also check power used each way.


come back with the numbers.

I would use  mid setting bitmain firmware  that gives about 13500 th and 1275 watts  94 watts a th

this guy should do 17000gh  and 1400 watts  = 82 watts a th  if you get this and then lose 2%  your eff would be 84 watts

you coin earned should be 34 x 0.00004156 x .98 = 0.0013847 in a day  using  24 x 2899 = 67.2 kwatts

vs 27 x 0.00004156 = 0.00112212 BTC in a day using 34 x 2700 = 61.2kwatts


at a 5 cent number  for power the new firmware  should mean you spend 3.36 a day and earn 4.91  a profit of 1.55 for 2 s9's

for bitmain middle speed  3.06 for power  earning 3.98  a profit of 92 cents for 2 s9's

Yes, you are right, you can not believe anyone. If someone takes the analysis of the firmware will be very grateful

Thanks for the review, but it would make sense coming for a trusted user who is not the OP.

not accusing for anything, but fake reviews are very common in the mining world.

therefore i advice everyone to make their own calculations and not to count on any other's.

thanks again.

I tested the firmware for 24 hours, Diffa# calculator showed 1.35%.
Poole took 15.4TH/s from 15.73TH/s, which is about 1.7%.
These figures do not exceed the declared Dev fee of 1.8%.
Electricity consumption 1480W at an ambient temperature of -2°C

The firmware takes into account the switching time between the pool so that at the output you get an honest 1.8% Dev fee

it really depends on the miner and the market condition, sometimes paying even 5%/10% of your profit  MAY still benefits you. for example the D3 by bitmain has a terrible firmware , therefore usually a custom firmware was needed , and the known custom firmware by then had less than 2% fees which was very reasonable and overall profit increased even after paying the fees. however in most cases the stock firmware does just fine, also one thing that many people seem to be missing is that when dev fees are stated at say 10% , it does not mean miner will mine for 2.4 hrs straight for the dev pool and then you get the all of the remaining 21.6 hrs to your pool, it doesn't work this way simply because

1- The dev fees are time-based, they start the count from the time the miner switches to their pool till the time it swihces back to yours, there is gap in between as it's never instant when cgminer switch pools , you are looking at a quite a bigger loss as you are the only who bears those switching periods, they are small but when summed up, the became reasonably questionable.

2- The dev ensures he get's his fees asap, so assuming it's 10% , they will try to get 10% of each hour, and not straight 2.4hrs. therefore assuming you start your miner, 2 mins on your pool , then 6 mins (10%) to dev pool , then cgminer restarts for whatever reason , 2 mins on your pool 6% to his pool again !

total time for you = 4 mins , for dev = 12 mins.

of course this case will not apply 100% of the time, but the possibility is always there, dev has the right to do it this way to prevent people from cheating,but they usually don't tell you the full story.
---------------------------------------------------

I tried a few custom firmware, i usually divide my miners , split them into 50% stock fw, 50% custom, let them run for a few days and them make my calculation.

i used the method mentioned above as well as Wireshark and other network tools to monitor exactly the time my miners were mining on dev pool vs mine.

there is almost 0 chance you are going to lose only what the dev states , there is always a bit to a lot extra .

so i advice everybody to do the math before trusting and settling down on a custom firmware.

That^^ plus the fact that IF the miner finds a block during the dev mining period you or the pool obviously do not get credit for the block -- the developer does. Just something to keep in mind...

you can simply monitor the hasrate on the pool side, a graph is usually presented. something like this



those ugly green arrows point to time periods that the dev fees were taking place. while this is not 100% accurate because a few other things aside form dev fees can show the same results ,but overall this is a highly accurate method to determine dev fees with little to no efforts.

take 24 hours of mining , sum the total "flat" time periods and get the percentage by doing a simple calculation.

do this for the first 24 hours if fees are higher than what you willing to pay > ditch it
if fees are fine , make a longer duration based test. maybe 72 hours. some devs are smart liars, they don't take the fees straight away,so you monitor the firmware for a few hours, and you say waw , this is it , only then they start to screw you , you just need to dig a bit deeper to be 100%.

the kernel log can also show you the results, but using the graph of your pool is usually the easiest way.

---

paying 7% for dev fees can be very stupid. simply because

assuming you went from 14 to 17th , then that's about 20% more hashrate, impressive right? but take a look at the power cost , it has increased, and you are paying 7% of your net-profit before paying for power. so at the end you might end up making just the same profit but risking your miner to failure by pushing it to it's limits.

Hello ,
concerning the firmware :
please put the date/time for each log file event .

Thanks
Regards

Make the calculation as in the picture, or measure per pool during the day

https://share.icloud.com/photos/0InWWkfkzHDdNPGfv7QeTjr6w

I really like the firmware, no complaints works very well. Job well done to dev team. Only problem is, and I'm no mathematician but dev fee seems much higher than 1.5%. Before taking steps to mass deploy, I'll need to figure that out.

Regarding Donat and how to calculate it. See the picture: https://ibb.co/yf4x7KL

how do you know it's 7%? any way to check this? 1.5~7% is a huge difference.

I'm using this fw and fee is 7% I think. But perfect fw for me!

Dev fee 1.8%, huh?

all works fine.

https://www.youtube.com/watch?v=Eq9WeXQlDwU&t=139s

Beware on using unknown firmware posted by newbies there are possibilities that the firmware is compromised.

Anyway, you can overclock and change the frequency without changing your firmware by following this guide https://bitcointalksearch.org/topic/how-to-overclock-your-autotune-antminer-s9-3239820
and increase the fan speed by following this guide https://bitcointalksearch.org/topic/solution-for-antminer-s9-fans-speed-up-down-every-20-seconds-2690140

its not working on my S9

Error when flashing --- "incorrect firmware"!!

Overclocking firmware antminer S9  17Th air cooling
20Th immersion cooling

Dev fee 1.8%

 -AsicBoost and LPM
 -Energy savings up to 20%
 -Autotune configuration
 -Overclock profiles
 -Disable Fans
 -Manual chips frequency config
 and much more

17.2Th  Power 1570W



Download firmware: http://bit.ly/2MPN3af