Yea i guessed that though i was wondering if there s an app that allows to relatively Simply change the P0 P5 state, reboot the cards, remove crossfire, etc.

Is there anything like qms that allows to modify all the tensions table easily for vega series on Windows with possibly latest drivers ?
As far as i understand, overdrivntool doesnt work for P1 P5 state for vega Under Windows ...

(i m new to vega 56, just have one rig and i m looking for an easy solution, i wont invest the same time as in RX series)
Many thanks in advance for your answer

Is there anything like qms that allows to modify all the tensions table easily for vega series on Windows with possibly latest drivers ?
As far as i understand, overdrivntool doesnt work for P1 P5 state for vega Under Windows ...

(i m new to vega 56, just have one rig and i m looking for an easy solution, i wont invest the same time as in RX series)
Many thanks in advance for your answer

First, you want to make sur that each board are well pluged to the uniting board, and then that the uniting board is well plugged with the orange Pi. It is the main cause of instability. Make sur to do this with G28 unplugged (not powered).

From my experience, when the previous was guaranteed ok, i had similar problems but with only with giant+ and baikal X, and it appears the first cause was the ethernert cable (replaced with better quality), or sd card neeeding reinstall (reformat or use a new sd card), or the orange pi needed to be swapped for a brand new one.

Hi,

I still have the same issue. Some are on one switch shared with Baikal-X (those ones, the Baikal X, runs at 99,9% Dacc).
Some others are on another switch and are the only one sharing that switch.


In both case they still are at around 90% Dacc.

What would you suggest that i test ? (and thanks a lot for your help !)

Thanks for your answer, i didnt had the time to follow on discord. I think too that sqrl has every Reason to be committed as long as their hardware is not vaporware. The market potential is huge if they achieve their goal (i know that in terms of Hashpower, it would be generally cheaper to buy new GPU to achieve same hahsrate, but in terms of powersavings, Acorn wins in the long run)

Guys i didnt follow on discord, do you mean SQRL has quitted the goal of programming a miner to use with their fpga ?

Thanks a lot Elder III

Hello everyone,

I went here out of curiousity and now I m excited again to make some tests (i built so many rigs for friends and me, so much that i stopped, bored to death, unless a friend really needs help). It s like i was discovering the RIG building and optimisation again 

I know its alpha, but i d like to buy a Nest2G card with 4 acorn 215 to test 2 different configuration (i guess the nest2g can drive all 4 GPUS plugged in USB using risers).

But of cours i have a lot of questions i dont know where to find answers :
- where can be downloaded the alpha mining program ?
- how is managed powered for the card when plugged in USB directly to nest2G (yes it will have the riser power and the direct power but wont it be missing the power it would have had from the PCIe where the riser is plugged ?


How to join a forum, community, source of information (i will do my job and search but i just need to be sure information is out somewhere i can access before ordering)


and yes i dont expect ROI from 4 accorns, it will be a drop in my glass of RIGS, but i expect test, failures, new challenges ... excitement in a worl that became dominated by "plug and play ASIC" (i dont blame them, i own quite a few)

Many thanks for your help and thanks for reading me

Guys, about DGB next algorithm to be Asic resistant (deployment wasannounced in january if i m right, after they said they tested it in november), do we know if they are targeting all Baikal DGB Algos ?
Also do we know if other coins like XVG have said anything about going the same way as DGB ?

Sorry to ask here but cant find any reliable information

I confirm the issue about assembly on 16 total G28 :
- multiple boards are bent, happily with no consequence for me
- needed to unmount and correctly mount fans because of their power cable being hit 
- due to transport, needed to replug the card linking the 7 boards, but on very few units, no damaged pins, using DHL. On the baikal X with FEDEX it was just horrible, bent pin on unmounted board ...


On skein or myr gr i dont have the crashing issue.


But on both on suprnova, zpool, miningpoolhub, for either DGB or XVG : i m missing 10% of rewards because i m missing 10% of hasrate compared to nominal 14 / 28. Whereas with BAIKAL X10 i m at 100% reward compared to nominal hashrate.

Those 10% are confirmed by miner stats : Dacc is around 89 90% but difficulty rejecty is only 0.1 or 0.2% ... where goes those 10 % ... neither the pool support nor baikal had an anwser ...

Hi,

Has anyone had both Z9 mini and innoslicon A9 exploited in same conditions (especially temperature) ? Which one is the quieter, is it by far or small difference ? Could you compare the noise of A9 to a Z9 mini set to 50% fan ?

(i ask because my noise limit in my environnement is close to Z9 mini at 50% fan).

Many thanks in advance

I really think 14GH is the capacity of G28 skein.
But there s Something wrong the "dacc" is around 80 to 90% on the G28 with difficulty managed by suprnova. This is too low, in comparison the X10 on same pool is around 99% explaining X10 is at nominal hashrate.

I tried to force difficulty on suprnova skein for the G28, but with no luck so far the best i achived is an average 12,8 GH/s which is still far from 14GH.

My guess is it is pool related. Though i contacted them twice by their support form and on twitter and NO answer since 48H.

Thanks a lot !!!!!!

That is exactly what i was hoping to find

Again thank you very much

Hi,

Is there a simple documentation explaining what option can be used in sgminer, especially for baikal but i would like to know complete list.

For exemple no-submit-stale (true or false) or api-allow (W xxxxx)

I checked on github but its the jungle for me.

Many thanks in advance for your help (you will do way better than frakin google on this matter)

Yes Groestl algo i have no problem.
Skein algo tested on suprnova, miningpoolhub and now zpool have all the same problems : hashrate betwenn -30% and -10%. Average on 12h on any of these pools is 11,8GH instead of 14GH

The option is baikal-fan
For Skein i set 10% and it remains below 30 degrees. It is still a little loud.
50% and 100% are almost as loud so i think thats why you hear no difference.

The main problem i have is the average hashrate reported on suprnova for G28 is 20% less than it should be (with high fluctuations it can goes from -40% to -10% averaged on 5 units). Whereas with X10 still on suprnova the average hashrate is 100% as advertised (fluctuations are small 5% with 5 units). this is for skein algo.
Anyone had the same problem ?

Ok to answer to myself :

Invalid shares was the result of either a faulty switch or IP conflict (all my switches being Netgear prosafe). Strange thing, if i add a new cpl plug i have the same colficting problem than whith the supposedly faulty switch …

In regards to Hashrate :
suprnova skein : its more around 13GH/s on average (calculated 2 hours)
zpool skein : i have the same result betwee 12,5 to 13,5 GH (2h with 5 x G28)

So far suprnova payout > by 20% of Zpool skein (paid in DGB), calcultaed on 24h with 5 G28 (not the one i used for tests so with stable hashrate). Luck difference being 5% in advantage to suprnova

I have the same problem about false hashrate.

Skein, advertise 14,19 GH/s but :
- suprnova : hashrate at around 12GH and 20% invalide share, those invalid shares show in Baikal UI too
- mining pool hub : hashrate average on 1 H between 11GH and 12GH, 12% invalid share on mining pool Dashboard and 0% invalid on baikal UI


Myr Groest : exactly same problem on suprnova (Verge).


I ll do more tests.

Nota : tests done with at least 5 x G28 unit at the same time

Hi, may i ask what are your settings and are they the same for Ethash and Cryptonight :

1/ in bios : you modded only memory straps ? if not what max core clock and voltage, same for P3 memory state

2/ other modifications either in bios (TDP for example) or using AB, wattool or other (mem clock etc.)

In my case my settings are made directly in bios after benching the card, generally i can always set max core clock at 1080 for 900mV, but i in too much cases i cannot set the memory clock to 2000 (worse i had is 1850). Strange thing though even at 1850 mem clock i can dual at 26MH + 60 MH lbry and i usually set TDP and max W at 110 in bios, resulting in a small economy of around 10 15w
Though my settings are bad for cryptonight (i do not car for now)