Question 3080 FE Afterburner undervolting not working

Jun 5, 2022
3
0
10
0
I have a 3080 FE on an open test bench, so cooling is not an issue. But I'm trying to make it quieter, because it's an FE :sweatsmile: However my power draw at a given voltage level seems really high. When looking at undervolting posts I should get around 240 watts at 800 mV. But my card still draws around 350 watts.
I was told not to use "Prefer Maximum Performance" in Nvidia control panel. But it's already set to "Normal" mode. And other modes are no longer available.

Do I misunderstand something or is this a technical issue? Do I have to configure Afterburner or Windows in a specific way for this to work properly?

Thank you!

 
Last edited:
Jun 5, 2022
3
0
10
0
Thanks, yeah I did exactly what that post said. Voltage control is enabled in "extended MSI" mode. Also tried "reference design", but no success.

I can get the power draw down to around 250 watts if I lower my core voltage to close to 700 mV. Could there be other things at play here, or even silicon quality? Could it be that the card was used for mining and that damaged it? If start a stress test from cold the initial power draw is around 310 watts @ 800 mV. Then it rises with temperature up to 350 ish.

I have already DDU'd and reinstalled the latest Nvidia driver. All other tools are on their latest version too.

 
Last edited:

Aeacus

Champion
Ambassador
Could it be that the card was used for mining and that damaged it?
If you have used GPU, with possibility that it was used for mining, then there's no telling what is wrong with it. But that much is sure, that mining will damage the GPU.

Consumer grade GPUs are not designed to run 24/7/365, with 100% load, almost at the level of thermal throttle. <- This will wear out the GPU - fast.

Steve talks further about it in here, at 11:42;

View: https://www.youtube.com/watch?v=g1Yp6pQRdns
 
Jun 5, 2022
3
0
10
0
Yeah, it's used. I heard about all the mining damage possibilities. But as Steve said it's mostly a "it works until it doesn't" scenario. But the behaviour of mine is odd.
I guess my question is, is this plausible behaviour for a mining damage? Or do I have some weird technical hickup and how could I possibly test for that?
 

Aeacus

Champion
Ambassador
I guess my question is, is this plausible behaviour for a mining damage?
It may not be mining damage, but it could be mining BIOS on GPU, which e.g set restrictions to all and any undervolting (and loss of revenue due to that). Flashing the BIOS back to default one, might fix it. Or when BIOS flash gets interrupted, for whatever reason, it will brick the GPU. Just like it is with MoBo and it's BIOS updates.
 

ASK THE COMMUNITY