Question I need help with my rtx 3080

[luisroani]

Firstly, I apologize for my weak English.

I recently purchased a used RTX 3080 10GB Aorus Master, and I'm having some problems with it.

When doing benchmark tests or even in games I notice that even though it is at 100% use, it is consuming less energy than it should, staying in the 60% range and because of this I have had low fps.

When running a furmark for example, in the first two seconds of the test it manages to pull the TDP to 90%+, but soon it drops to 60% and drops until it reaches 50% of energy consumption, although it is at 100% of its use.

I've tried several things like:

Test Windows 11

Change the bios switch to oc/silent on the video card

Test on another MB and PSU (showed the same defects)

I honestly don't know what else I can do

Below is a screenshot of it on the furmark after a few seconds of testing.


how to see at max fps, in the initial seconds it reaches 451 fps, after that it reaches a mediocre 200-240fps

https://imgur.com/a/lsTJouz

 

[Lutfij]

Welcome to the forums, newcomer!

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

I recently purchased a used RTX 3080 10GB Aorus Master, and I'm having some problems with it.
I'm assuming you had a discrete GPU prior to the RTX3080. If so, did you run DDU and remove all GPU drivers(Intel, AMD and Nvidia) then manually reinstall the latest GPU driver in an elevated command, i.e, Right click installer>Run as Administrator.

 

[luisroani]

Sorry i new here, all componets are new except for gpu

CPU: i5 12400f
Cooler da CPU:Stock
Placa-mãe: Duex dx-h610zg
Ram: 16GB Single Channel Fury 3200mHZ DDR4
SSD/HDD: 512GB western digital nvme
GPU: RTX 3080 Aorus master 10GB Rev 1.0
PSU: Azza 750w Bronze
Chassi:
SO: W10
Monitor: Rog Strix pg279q

I can't say what the BIOS version of the motherboard is because if I want to find drivers for it as it is a very simple board, probably a relabel from a Chinese brand.

However, I already tested it on another Gigabyte motherboard and had the same problem.

 

[Phaaze88]

Download and run Gpu-Z, then click Advanced tab.
In the drop-down box, select Nvidia bios. Post that image here.
The info there will help identify if a mining bios is in use.


Also:
-the single dimm of ram will hurt performance. That will slow down how much information gets fed to the gpu.
-that psu is possibly harmful to the gpu, which is throttling to protect itself from it. [I tried looking it up in PCpartpicker, but got nothing.]
-Is a return possible?

 

[luisroani]

Thank you very much for your answer:

A return is not possible

I know that just one RAM slot can decrease performance, but what I'm experiencing is a decrease of half the expected performance.

The power supply is really questionable, but I've already tested it on other gpus of similar power and none of them had this defect.

Below is a link with a photo of the Nvidia bios data on gpu-z

https://imgur.com/a/FA2WdX5

View: https://imgur.com/a/FA2WdX5

 

[JohnBonhamsGhost]

recently purchased a used RTX 3080

from where did you purchase

 

[luisroani]

from where did you purchase

I bought it from a guy who had advertised on the Facebook marketplace
As I said, the video card is used, and it is difficult to say what use was given to it previously.

What I can report is that as soon as I open a benchmark it seems to pull the ideal power of 300-350w for 1-2 seconds, after that it drops to 200w.

in games I notice that the card usage is 50-60% most of the time even without any frame synchronization

 

[Phaaze88]

https://www.techpowerup.com/vgabios/232185/gigabyte-rtx3080-10240-210315-3
F4 in Gigabyte's website.
So the vbios isn't a mining one, so that rules this one out.

The power supply is really questionable, but I've already tested it on other gpus of similar power and none of them had this defect.

Other 3080s? Perhaps a 3090?

 

[luisroani]

Yes, I also used other models of the 3080 (gainward) with the same power supply, I also do tests with the 3080 aorus master on an EVGA 700w bronze power supply and I have the same problem, the TDP barely reaches 50%

https://www.techpowerup.com/vgabios/232185/gigabyte-rtx3080-10240-210315-3
F4 in Gigabyte's website.
So the vbios isn't a mining one, so that rules this one out.



Other 3080s? Perhaps a 3090?

 

[luisroani]

What surprises me is the fact that at furmark in the first second of testing the card reaches the ideal TDP of 90-95%, and soon after it suddenly drops to 50%, could the temperature be related to this? because the card varies between 79c-81c in the furmark test and gpu-z reports a hot stop of 104c

 

[Phaaze88]

hot stop of 104c

Hot spot shouldn't cross 100C, but does it do that in an actual game? Furmark is a bit of an exception, as it's a constantly high load, and games aren't really like that.

 

[luisroani]

Yes, in games the hotstop is also between 100-104c, at 100% use of the card, however the fps delivery is not even close to what was expected

 

[JohnBonhamsGhost]

in games the hotstop is also between 100-104c

this would be causing throttling and is likely why the power % drops so quickly when in use.

I bought it from a guy who had advertised on the Facebook marketplace

this is probably why it was being sold.
hopefully you didn't pay much for it.

i would probably contact the seller and see if they would offer some type of refund or exchange.
could turn out to be a decent person, but i wouldn't expect it if they sold this with no mention of issue.

 

[Phaaze88]

Yes, in games the hotstop is also between 100-104c, at 100% use of the card, however the fps delivery is not even close to what was expected

Since there really is no way to return this - which sucks - and get your money back, you're stuck trying to fix it, or pass it off to someone else...

As for how to fix it...
This old HWUB review suggests the cooler chokes itself somewhat; that the shroud covers too much of the heatsink.
A repaste with some NT-H2 or MX-4 may help shrink the gap between gpu core and hot spot, but there's still the bloody shroud to contend with.

What is your case? That's going to influence the airflow around the cooler too.

 

[luisroani]

I spoke to the seller just now, he informed me that I should change the thermal paste and add thermal pads to the card, or else he gave me the option of returning it, as I have no experience in opening a video card and making these changes, I opted for the return, thx for the help