GPU driver crashes a lot lately

[HeLiOn]

Hello.
I have a Gainward GeForce 4080 Super JetStream OC running on an Asus TUF GAMING B650-Plus.
For about a week now, the driver has started to randomly crash. The screens go black, and the fans start going at max speed.
At first I thought it was a driver issue, so I reinstalled the driver. It didn't solve it.
2 Days ago it crashed 5 or 6 times, so I tried removing the driver completely using DDU, and installing an older driver version from December (566.36).
That didn't solve it either. Yesterday it also went out like 5 or 6 times.
Then I booted my PC on Linux to see if it's an OS issue. Fired up Metro Exodus, but after the first game restart (after doing the settings) the driver crashed again.

I'm starting to have cold sweats here, cause if this is a hardware issue, and I have to send it back, the store will just give me my money back, and I won't find a 4080 Super again.
Is there anything else I can try here, before giving up on this card completely?
if it's of any relevance, I'm not using HDR or any form of variable refresh rate.

 

[Mumintroll]

You could try the 4080 in another pc. Bring it over to a friend to test if you don't have a spare pc.

Also test another gpu in your pc if you can.

[Mark Kaine]

How do you know it's the driver? or even the gpu at all?

 

Full specs ?

[HeLiOn]

46 minutes ago, Mark Kaine said:

How do you know it's the driver? or even the gpu at all?

 

Full specs ?

They're hidden in my signature

[Tan3l6]

How about the PSU?

[Mark Kaine]

1 minute ago, HeLiOn said:

They're hidden in my signature

Ok, I see. Have you tried no expo/XMP? 

 

And, again I gotta ask.

1 hour ago, HeLiOn said:

the driver crashed again

What's the error message?

[HeLiOn]

1 hour ago, Mark Kaine said:

Ok, I see. Have you tried no expo/XMP? 

I got that suggestion from discord earlier.
I did check the XMP profile and it was already disabled.
I don't know for how long it has been like this, so for now I enabled it and I'll see if I get anymore crashes.

 

1 hour ago, Mark Kaine said:

And, again I gotta ask.

What's the error message?

I don't know of any error message. When the crash occurs, my screens go black (they stop receiving input), and the fan speed maxes out.
If there is any error message, I'm not able to see it.

[HeLiOn]

1 hour ago, Tan3l6 said:

How about the PSU?

It's a Be Quiet Dark Power Pro 12, 80+ Titanium, 1200W.
I'll add it to the specs.

2 hours ago, Mumintroll said:

You could try the 4080 in another pc. Bring it over to a friend to test if you don't have a spare pc.

Also test another gpu in your pc if you can.

If everything else fails, I'll try that.

[HeLiOn]

2 hours ago, Mark Kaine said:

Ok, I see. Have you tried no expo/XMP? 

I can now confirm that this happens regardless of whether XMP is active or not.

[Mark Kaine]

2 hours ago, HeLiOn said:

I don't know of any error message. When the crash occurs, my screens go black (they stop receiving input), and the fan speed maxes out.
If there is any error message, I'm not able to see it.

That's why I'm asking because you can't know at all if it's the GPU if there isn't a specific error message and even then it's not always clear...

From your description this is either RAM, motherboard (BIOS) or some kind of software issue.

 

Is latest BIOS installed?

Did you try DDU?

ran memtest64?

 

Also check event viewer for "critical" errors.

Especially WHEA errors but also everything else can be important if it's "critical".

 

 

[HeLiOn]

3 hours ago, Mark Kaine said:

That's why I'm asking because you can't know at all if it's the GPU if there isn't a specific error message and even then it's not always clear...

From your description this is either RAM, motherboard (BIOS) or some kind of software issue.

 

Is latest BIOS installed?

Did you try DDU?

ran memtest64?

This thing occurs on both Linux and Windows.
That's why I assumed it's either a hardware issue, or something wrong at a driver level.
Or some weird stuff between the motherboard and the GPU...

I have the latest bios installed. I did the upgrade 3 days ago in an attempt to solve this issue.
Also, I did try DDU. As mentioned in the OP I used it to completely remove the latest Nvidia driver and then I installed an older version from December.
I now downloaded memtest64 and ran 2 loops. I honestly don't know if I did this one right, but here's a screenshot with the results:
 

Spoiler

 

3 hours ago, Mark Kaine said:

Also check event viewer for "critical" errors.

Especially WHEA errors but also everything else can be important if it's "critical".

I'm going to need some help with the event viewer. I have no idea how to navigate that thing.

[Mark Kaine]

1 hour ago, HeLiOn said:

This thing occurs on both Linux and Windows.
That's why I assumed it's either a hardware issue, or something wrong at a driver level.
Or some weird stuff between the motherboard and the GPU...

I have the latest bios installed. I did the upgrade 3 days ago in an attempt to solve this issue.
Also, I did try DDU. As mentioned in the OP I used it to completely remove the latest Nvidia driver and then I installed an older version from December.
I now downloaded memtest64 and ran 2 loops. I honestly don't know if I did this one right, but here's a screenshot with the results:

Keep in mind I didn't mean it can't be the GPU, it just doesn't sound like it to me... Did you run some benchmarks? I recommend 3dmark FIRESTRIKE... it definitely tends to crash if "something" is wrong, worth a try and might be interesting.

 

As for event viewer it's super easy... Just open it, first category is overview...

 

 

If there are any critical errors just click on them and post a screenshot of the error descriptions.

[Mark Kaine]

2 hours ago, HeLiOn said:

I honestly don't know if I did this one right,

My bad, I meant memtest86, you gotta boot it from USB... (memtest64 isn't as good) But it's a shot in the dark anyways, it's entirely possible to not fail the the test and still have bad RAM... As it's often a compatibility issue and the RAM may not be defective at all (and that's what memtest tests...) 

 

https://www.memtest86.com/

 

 

[HeLiOn]

2 hours ago, Mark Kaine said:

Keep in mind I didn't mean it can't be the GPU, it just doesn't sound like it to me... Did you run some benchmarks? I recommend 3dmark FIRESTRIKE... it definitely tends to crash if "something" is wrong, worth a try and might be interesting.

 

As for event viewer it's super easy... Just open it, first category is overview...

If there are any critical errors just click on them and post a screenshot of the error descriptions.

The event viewer only shows the forced restarts I have to do, in order to get the PC running again.
This last crash I had, I tried to see if I can still operate the PC through Teamviewer, but when the crash occurs, Teamviewer also stops receiving video signal.

[xg32]

10 hours ago, HeLiOn said:

.

use old drivers that are 566.xx, do not use any of the 57x.xx drivers, they have power issues imho. if it persists then i hope it's something else and not permanent damage.

 

I'm still waiting to see if my 3080ti is cooked, already looking for a new gpu near msrp.

[HeLiOn]

1 hour ago, xg32 said:

use old drivers that are 566.xx, do not use any of the 57x.xx drivers, they have power issues imho. if it persists then i hope it's something else and not permanent damage.

I did try that already. Mentioned it in the original post.
 

4 hours ago, Mark Kaine said:

My bad, I meant memtest86, you gotta boot it from USB... (memtest64 isn't as good) But it's a shot in the dark anyways, it's entirely possible to not fail the the test and still have bad RAM... As it's often a compatibility issue and the RAM may not be defective at all (and that's what memtest tests...) 

https://www.memtest86.com/

I performed a memory test and it passed.

These are the results:
 

Spoiler

 

4 hours ago, Mark Kaine said:

Did you run some benchmarks? I recommend 3dmark FIRESTRIKE... it definitely tends to crash if "something" is wrong, worth a try and might be interesting

 

I can try, but I'm not sure if they'll be relevant here.
This crash is completely random.
Sometimes I get to game for hours and nothing happens, just for it to crash when simply browsing the web.
 

[Mark Kaine]

18 hours ago, HeLiOn said:

The event viewer only shows the forced restarts I have to do, in order to get the PC running again.
This last crash I had, I tried to see if I can still operate the PC through Teamviewer, but when the crash occurs, Teamviewer also stops receiving video signal.

Tbh... It's probably RAM... I know you won't believe that - but after 100s.... 1000s...? Of very similar cases.... It's either RAM or some driver issues (but since there's nothing in event viewer it's just way more likely RAM... sometimes there can be WHEA errors, but not even that)

 

A BIOS update could fix it, new ram could, new motherboard, etc, etc...

 

Of course it could be GPU or CPU as well... It's just very - very unlikely (since nothing really points to that) random crashes with no relevant traces? Yeah, it's gotta probably be the ram/memory controller......

 

You probably think ram "just works" but it's far from it...

 

IF you get new ram, do some research what other people use with your mobo etc.

 

Alternatively: just start swapping parts until you find the culprit, aka traditional troubleshooting (I just know I'd start with ram lol)

[Mark Kaine]

PS: what's your ram? You check it's on the QVL?

 

https://www.asus.com/motherboards-components/motherboards/tuf-gaming/tuf-gaming-b650-plus/helpdesk_qvl_memory?model2Name=TUF-GAMING-B650-PLUS

[HeLiOn]

5 hours ago, Mark Kaine said:

You probably think ram "just works" but it's far from it...

I have enough experience to know better.

And I did careful research before I even bought these parts.
I know, a lot of people get these things wrong, so I'd question this too.

5 hours ago, Mark Kaine said:

PS: what's your ram? You check it's on the QVL?

https://www.asus.com/motherboards-components/motherboards/tuf-gaming/tuf-gaming-b650-plus/helpdesk_qvl_memory?model2Name=TUF-GAMING-B650-PLUS

My specs are hidden in my signature. You just need to expand it.
And yes, the RAM I have are in there:

 

Since I couldn't afford to leave the PC like this until tomorrow (it would affect my work), I swapped GPUs with an older PC, like @Mumintroll suggested.
I got my old 2080 Super in here now and the 4080 Super is now on a Gigabyte Aorus Gaming 7 motherboard along with an Intel i7 8700K.
I did the swap around 4 hours ago. Since then, that PC has stayed open, and I even went through the prologue of The Last of Us Part One, on 4K max settings.
So far, there hasn't been an incident on any of the rigs, but I'll keep things like this for another one or two days, and see if any the setups acts up.

For now I'll just keep the thread open and come back with an update.

[HeLiOn]

I can finally rule out this being a hardware issue (and sleep better at night).
Ever since I swapped the GPUs, I haven't had a single incident on any of the rigs.
So yesterday. I finally swapped them back, and so far, everything has been going fine.

Since these stopped happening after the first swap, I think I'll just pass it on the GPU being somehow improperly seated.
One of the PCI connections must have been broken, I don't know...
I'll post if I find something else, but for now I'll just enjoy the smooth sailing.
Thank you for the help, guys...