©ASUSTEK COMPUTER INC. ALL RIGHTS RESERVED.

Krorghar · ‎11-21-2020

Just installed the card, a TUF 3080 OC version, and begun my testings. All other games seems to work fine, like Rise of the Tomb Raider, and Mafia Definitive edition, But witcher 3 resets the comp as soon it has finished loading, is not a CTD, is a hard reset that makes the comp reboot without warning. The TUF replaces a GTX 1080 that had no previous problems in running witcher 3 at max detail, bit low on fps, but that was all.

The card is running as it, out of the box, no aditional OC Done, Drivers are nvidia, last version, Downloaded the BIOS utility, and it says card needs no update. And the PSU is a seasonic 1050W.

Any clue guys, or I have just got a defective card and should RMA it?

Loranon · ‎11-21-2020

A sudden reboot is more of a sign for a PSU problem than a GPU problem. Did you properly connect all PCIe power plugs?

As the 1080 draws less power than a 3080, it could be that you just didn't notice before. Maybe the PSU has a bad rail (don't know your exact model).

You could try if lowering the power target via Afterburner mitigates the issue and if so try to identify what exactly causes it. Switch PCIe cables, reseat GPU as well as power connectors, etc.

Krorghar · ‎11-21-2020

Ty Loranon for the response.

Installed Afterburner, and lowered the power slide to 50% and witcher 3 remained stable the only try i did, this triggered all alarms. Installed GPU-Z to monitor card power consumption and this is what i got.

Dont know if these power numbers are normal, but thats with me sitting at desktop looking at the monitor, I maybe wrong but it seems the card board is sucking allmost 100W, the chip 17W, MVDCC 58W (dont know what this part is) and PWR_SRC another 65W (also dont know what part it is). If someone could explain these values, and tell me if its all normal for an idle TUF RTX 3080, would be very gratefull

Logged this data to a file, and in the moment of a crash the last 4 values logged where 283, 144, 73 and 88. This makes me think i may have a faulty card and is time to RMA.

BTW my PSU is a seasonic snowsilent 1050 platinum 3 years old. Got another brand new PSU I could swap and see if it dont crash, but first would love to know if my card is all ok, or these numbers are too much, as reviews says it should be sucking 20W while idle.

Loranon · ‎11-22-2020

Almost 100W in idle does indeed seem odd. It also seems the card is not clocking down properly and instead sits at 1700ish Mhz in idle also odd. For reference: my 3080 Strix OC sits at 38W idle at 210Mhz.

Now this may have various reasons, nVidia control panel "performance" mode is known to prohibit the card from clocking down. Try changing to "optimized" power mode. However that would not explain the crashes.

In any case, this is not normal behavior. However, 280W is, as far as I know, quite a usual power consumption for the card. I would try a few more things on the software side, including a driver reinstallation as well as reseating and reconnecting all cables. However, it does seem unlikely, that the connectors are the issue here.

And since you already have the PSU at hand, I would try to replace the PSU before going the RMA route, just to make sure it really is the card.

Krorghar · ‎11-22-2020

Many thanks loranon, this afternoon will replace the cables and the PSU and check it, but if the card keeps consuming 100W idle, its a problem and I think it will go RMA

Have changed power setting to optimized, and still got the same consuming, 😞

BTW got some question about the data pls,

Board power draw, is the total power draw of the full card?
GPU Chip power draw is self explanatory, and I presume MVDCC should be the power draw by the VRMs but what is PWR_SRC?
Should the sum of these 3 values be the same that total board power draw?

Loranon · ‎11-22-2020

I am not entirely sure what SRC_PWR means in the context here. I actually assume that GPU-Z shows invalid, or just plain wrong, values. As far as I know, there is not such value to read from Ampere cards, hence I don't know where GPU-Z gets the value.

The "relevant" value is, as you mentioned, the board power draw, as it is roughly the clean power output of your PSU. Note that the actual power draw from the wall socket will be slightly higher due to losses.

Good luck changing the PSU - I hope you can resolve your issue!

Krorghar · ‎11-22-2020

Wel loranon this is weird m8.

As first step, I Swapped the 2 old, 8 pin cable from my old PSU for brand new ones from my reserve PSU, and problem solved. I dont know what was causing the problem at first, If it was a bad slot conection, a defective cable, or a bad rail in my PSU (i swapped slots too).

Now there was a weird fact. In the begining the system had two GTX 1080s. As i was going to get the RTX 3080, I gave one of the 1080s to my brother for his PC, but as the power cables were secured with nylon brides, and routed in the back of the case, i decided not to remove them, so they remained connected to the PSU. Now I suspect that since the cables were still connected to the PSU, it may have been sending power to them even when they were not connected to anything. This makes no sense, as a cable connected to nothing, should not have power, but cant think on anything else.

Well many thanks for the help m8

Tuf 3080 OC Witcher 3 resets the comp