I am not sure how I can do that.
I noticed first GPU gets max 272W if go over that, hash suddenly drops to 40-50 then increasing again
From the hiveos.farm web interface:
- find the old school command prompt icon
- select Hive Shell Start (if remote LAN)
In the menu area below the ribbon but above your Rig/GPUs status screen, you’ll see “hive shell” activity and a little white box will show up…click this box and a new window opens.
At that new remote console…
enter: Nvidia-info at the prompt
1 Like
1 Like
You found the Throttle: PowerLimit on each GPU, right?
Note: in this case 2 more commands may be helpful
- nvtool --throttle can also be used to keep the output more condensed/focused
Comprehensive tool
- nvidia-smi -q -d PERFORMANCE
At some point, it is there and may be contributing to the performance.
Note: I have mine set at 305 even though it “reports” far less.
1 Like
After changing settings on lhr cards you need to restart the miner, otherwise it will think lhr lock engaged since the clocks changed and attempt to unlock.
Check the 3080Ti hashrate boost using evga bios:
do at your own risk and blame no one if things go wrong.
I have several Palit Gamerock 3080Ti’s and after this bios update I was able to get 88-91 MH/s on ETH with 280 Watt PL. Core: 1200 Mem: 2400 on latest HiveOS.
Also, I have palit gaming pro 3080Ti’s and somehow this hack did not work on gamingpro’s.
Good luck.
1 Like
Thank you so much, thats a life saver video, cards are stable now, no issue at all, installed this evga bios to my all cards, all worked great.
S O L V E D…
1 Like
What bios flasher did you use?
Interested if you followed any tutorial.
I followed him exactly, but please take note;
After mine a while I noticed that temperatures hotter than before,
And there is sudden WATT changes, jumping from 100 WATT to 280 WATT directly at the first start or after you changed your overclock settings, not sure if that is okay.
1 Like
To achieve more MH try to:
- remove power limit
- set core clock to 1140
- set memory clock to 2500 (i’m using 1250 on windows and it’s super stable)
1 Like
Yes, this is pretty ok.
Temperature are higher because now gpus work properly so running at maximum MH you can mems heat more (and stock pads usually sucks)
Wattage are unstable on EVERY lhr model gpu because the lhr indeed, so don’t worry. If you have a wall watt meter you can see yourself. Not only when applying overclock but ALL THE TIME power usage fluctuates a lot, that’s ok
1 Like
Maximum I can go 2200 memory, after that getting GPU no temps error.
Currently its stable with:
1200 2200 300
Yeah mate, knowing people getting 100 mh with 220watts while I am getting 87 MS with 300watts so annoying, hope nvidia get covid, don’t know what to say.
Actually 3080ti is known to mine (at 74-75%) about 90mh, can’t hit 100mh at all, unless you got Dell Alienware one which is basically 3090 but this is an other story actually.
And yes again, this is pulling more power because it’s supposed to mine at almost 120mh so why those 300w power coming out.
1 Like
And about this point ok, so that’s your limit, this is the silicion lottery as well, i can keep mining with no problems my 3080ti at 1600 (3200 on linux) as well but with no improvements in terms of hashrate so i found 1250-1300 is my limit where mh do not increase more.
1 Like
albyG, thanks for informations mate, can I ask you if you have a solution about this issue, any level above 2200 memory clock causes GPU driver error, no temps. Is there a way to fix that or that’s that card maximum can go for 2200?
2 friends of mine who have 3080ti both have very very different OC, one can easy hit 3000, the other one barely touch 2200, after that he get same errors you get, so crashes, gpu driver errors and so on.
I also have right the same problem on my 3070ti which is very fragile, and more than 2200 cause crashes. One of those 2 friends of mine also have 3070ti and he mine at 2600 on memory.
This is just the classic silicion lottery unfortunately you can’t do anything about that, it’s just unfortune.
Try running your at 2100 on memory and look at mh it produces, if you are in range of 88-91 mh you are ok, otherwise increase memory clock by 25-point-steps. That’s all.
One good thing to do is the replace thermal pads, but those do not actually compromise the hashpower, at least at the very beginning
1 Like
Thank you so much mate.
hi! I have 2 pny boards with exactly the same problem as yours, will that bios mod work? error gpu driver error no temp, and max hash rate 74