RTX 3000 Series artifacts, black squares that flicker in desktop and 2D mode

[erocker]

These are pretty classic vram artefacts. The card is defective and needs to be replaced.

 

[eidairaman1]

These are pretty classic vram artefacts. The card is defective and needs to be replaced.

Whats with all these rtx 3000 failures, reminds me of the rtx 2000 debacle

 

[Dr. Dro]

I own an EVGA 3090 and have consistently been dealing with this issue since day one. I've tried everything except a new monitor. Some drivers seem to work better than others but it seems like the issue gradually gets worse over time. Ironically, mine acts up when the PC has been idling for a while and then I go to game on it. A reboot always solves it for the gaming session.

This is not a memory clock issue. I have no doubts that this is a driver related issue. N India doesn't have a good track record with tracking down and fixing interment issues. I get they are hard to identify the root cause but nobody should be dealing with this after paying for a $1500 GPU!

I dunno about that. Could be that you got really unlucky with your setup, the only way I have ever been able to reproduce this problem is by overclocking the GPU memory too much.

Since you seem to be very prone to the problem and I have a hunch it has something to do with the card powering down after a gaming session with unstable VRAM settings, can you try your usual settings but with P-states disabled, make a note of the behavior?

Make a registry file and add this, replacing GUID with the correct entry for your GPU driver (GPU-Z can show this for you under Registry Path):

Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{GUID}\0000]
"DisableDynamicPstate"=dword:00000001

then reboot. Your card will no longer downclock from its maximum performance state anymore unless it hits power or thermal limits.

To revert, just delete the DisableDynamicPstate registry key and reboot again.

 

[Anomalistics]

These are pretty classic vram artefacts. The card is defective and needs to be replaced.

Have you read all of the replies? I even reached out to people on Discord that have had this issue as well, all very much the same story. Cannot replicate it, happens at idle, no issues on high loads, etc, etc.

 

[erocker]

Have you read all of the replies? I even reached out to people on Discord that have had this issue as well, all very much the same story. Cannot replicate it, happens at idle, no issues on high loads, etc, etc.

Yes I have. Same answer.

 

[pf100]

I know it happened with another card and not in games, but try underclocking the vram by 100mhz.

 

[ThirtyNinety]

Already running now for 3 weeks without seeing the problem again. "Card is defective" in this context is just bullshit, only said by complete noobs.

As it is certainly triggered by some software related to the GPU, or just the Nividia kernel driver conflicting with one of the startup programs sometimes, I did some important changes in my system startup including removing Brother Printer software autolaunching, removing Asus Armoury Crate with the uninstall tool and also disabling the "vCore Downgrade" setting and the "Asus DIP Away mode" in Asus AI Suite and scheduled tasks.

I dunno about that. Could be that you got really unlucky with your setup, the only way I have ever been able to reproduce this problem is by overclocking the GPU memory too much.

Since you seem to be very prone to the problem and I have a hunch it has something to do with the card powering down after a gaming session with unstable VRAM settings, can you try your usual settings but with P-states disabled, make a note of the behavior?

Make a registry file and add this, replacing GUID with the correct entry for your GPU driver (GPU-Z can show this for you under Registry Path):

Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{GUID}\0000]
"DisableDynamicPstate"=dword:00000001

then reboot. Your card will no longer downclock from its maximum performance state anymore unless it hits power or thermal limits.

To revert, just delete the DisableDynamicPstate registry key and reboot again.

Did you test this recently? I tested this for my 3090 and it didn't work with the latest Nvidia drivers (yes rebooted etc)

 

[Dr. Dro]

Did you test this recently? I tested this for my 3090 and it didn't work with the latest Nvidia drivers (yes rebooted etc)

I always run that key as I do not like dynamic GPU clocks, but I have not had the black squares in a very long time.

If it doesn't work for you, it's caused by Windows... it keeps GPU driver settings stored in a fast reload location nowadays, even if fast startup is off.

You can ensure that your computer will always reload all settings fresh upon reboot by doing it from the command prompt:

shutdown -r -t 0

 

[ThirtyNinety]

I always run that key as I do not like dynamic GPU clocks, but I have not had the black squares in a very long time.

If it doesn't work for you, it's caused by Windows... it keeps GPU driver settings stored in a fast reload location nowadays, even if fast startup is off.

You can ensure that your computer will always reload all settings fresh upon reboot by doing it from the command prompt:

shutdown -r -t 0

Hmm not needed to reboot imo as you can just do this with Win+Ctrl+Shift+B
Actually, the problem people describe here and the fix they aim for that is about not having to reboot anyway. Also, I had the issue only on cold boot, with rebooting it was always gone without using this legacy command prompt which imo does not do anything special. Even in BIOS the squares were already gone, which proves that this is a software issue.

 

[Fluffmeister]

Have you read all of the replies? I even reached out to people on Discord that have had this issue as well, all very much the same story. Cannot replicate it, happens at idle, no issues on high loads, etc, etc.

Agreed, had it once... completely randomly from a cold boot, black squares on the desktop, rainbow effects in Mordhau on the same boot cycle. Never had the issue before or since.

 

[Dr. Dro]

Hmm not needed to reboot imo as you can just do this with Win+Ctrl+Shift+B
Actually, the problem people describe here and the fix they aim for that is about not having to reboot anyway. Also, I had the issue only on cold boot, with rebooting it was always gone without using this legacy command prompt which imo does not do anything special. Even in BIOS the squares were already gone, which proves that this is a software issue.

That command issues a reset command to the driver while it is still loaded, but it does not work, in fact, I don't think that command has ever worked correctly. Fully unloading and reloading the driver through either device manager or using the CRU monitor utility's restart64.exe file work much better when you are dealing with such lower-level settings.

It's the VRAM settings being too high to be stable. That's about it. BIOS doesn't use full power state, so the problem would never manifest itself there unless your GPU had hardware damage in it.

 

[ThirtyNinety]

That command issues a reset command to the driver while it is still loaded, but it does not work, in fact, I don't think that command has ever worked correctly. Fully unloading and reloading the driver through either device manager or using the CRU monitor utility's restart64.exe file work much better when you are dealing with such lower-level settings.

It's the VRAM settings being too high to be stable. That's about it. BIOS doesn't use full power state, so the problem would never manifest itself there unless your GPU had hardware damage in it.

This command does restart the driver here but gives indeed a filthy Windows error event id 4101. Anyway, it's irrelevant here, was example and it should work out of the Nvidia box without the need to restart.
Can you elaborate VRAM Settings? I suppose you mean the VRAM Clock here, as I don't think there is a lot to be set as the driver decides this independently, this in the case when you don't use tools like MSI Afterburner or GPU Tweak of course.

 

[Dr. Dro]

This command does restart the driver here but gives indeed a filthy Windows error event id 4101. Anyway, it's irrelevant here, was example and it should work out of the Nvidia box without the need to restart.
Can you elaborate VRAM Settings? I suppose you mean the VRAM Clock here, as I don't think there is a lot to be set as the driver decides this independently, this in the case when you don't use tools like MSI Afterburner or GPU Tweak of course.

VRAM clock, yes. It's too high for whatever workload you're doing, and producing errors which accumulate and once the frequency falls, something goes wrong with the error correction and produces said artifacts. Needless to say this should not occur with stock clocks on VRAM, if it does, then you likely have a hardware fault somewhere. Possibly thermals if card is fine otherwise, 3090s are known to be exceptionally hot on the memory side.

Driver will only change power state under full load in compute workloads, going from P0 to P2 which is usually -250 MHz or so, i.e. from stock 9751 to 9501 reading. This still occurs with overclocked memory, i.e. mine goes from 10552 to 10302, as I understand, this is intended behavior on GeForce.

 

[ThirtyNinety]

VRAM clock, yes. It's too high for whatever workload you're doing, and producing errors which accumulate and once the frequency falls, something goes wrong with the error correction and produces said artifacts. Needless to say this should not occur with stock clocks on VRAM, if it does, then you likely have a hardware fault somewhere. Possibly thermals if card is fine otherwise, 3090s are known to be exceptionally hot on the memory side.

Driver will only change power state under full load in compute workloads, going from P0 to P2 which is usually -250 MHz or so, i.e. from stock 9751 to 9501 reading. This still occurs with overclocked memory, i.e. mine goes from 10552 to 10302, as I understand, this is intended behavior on GeForce.

What you are saying would sure make sense, however on one condition, namely if we were not here talking about a very particular problem here: every now and then, and ONLY when after boot, we get these artifacts. So there is no previous workload. There were no games which were played/closed previously or any "accumulation of errors". I guess nobody really reads the descriptions of the problem here anymore, at max just looks at the movies in the initial posts, and based on that has already made a final conclusion, without knowing the specific context in which this issue is generated.

 

[Dr. Dro]

What you are saying would sure make sense, however on one condition, namely if we were not here talking about a very particular problem here: every now and then, and ONLY when after boot, we get these artifacts. So there is no previous workload. There were no games which were played/closed previously or any "accumulation of errors". I guess nobody really reads the descriptions of the problem here anymore, at max just looks at the movies in the initial posts, and based on that has already made a final conclusion, without knowing the specific context in which this issue is generated.

Oof, I have never had this problem from a cold boot, if I had, I would have seriously freaked out. You could perhaps try to have MSI Afterburner apply a very aggressive negative offset to frequency (just set it to -1000) on-boot for debugging purposes, but it is looking like your card needs an RMA.

 

[ThirtyNinety]

Oof, I have never had this problem from a cold boot, if I had, I would have seriously freaked out. You could perhaps try to have MSI Afterburner apply a very aggressive negative offset to frequency (just set it to -1000) on-boot for debugging purposes, but it is looking like your card needs an RMA.

But no, that just proves the whole point in me saying that lot of you did not read a lot of the message flow above here:

- Most of us already did at least one RMA already, or bought a new GPU (me f.i. replaced my 3080 for a 3090 and had exact the same issue)
- Currently I am not even having this problem anymore now for more than 3 weeks. This after doing massive changes on my system (including configuring less aggressive Windows TDR delay, which I see now I actually forgot to mention lol)

 

[jesseinsf]

I have this exact same issue. For me, it comes intermittently when starting up my PC. It will show on the lock Screen, sign-in screen, and desktop. To get rid of it, all I need to do is reboot my system. An issue that also comes with this experience is that I cannot access the system BIOS when rebooting my PC where the monitor turns off before system BIOS POST has had a chance to complete. The monitor stays off until the lock Screen shows (The newest VBIOS that is supposed to fix a similar display issue DOES NOT WORK). Plus, the BIOS LED light stays white when it is supposed to be off. I have replaced this Video card twice and both have the same issues on multiple monitors and while using multiple DisplayPort cables. The latter issue doesn't happen when I turn on my PC or when the BIOS posts before the monitor turns off during a reboot.

See my system specs for details of what I have.

• System Name: Windows 11 Pro for Workstations Build
• Processor: Intel Core i9 9900K
• Motherboard: ASUS ROG MAXIMUS XI EXTREME
• Cooling: ASUS ROG STRIX LC II 360 ARGB AIO, 10x UNI FAN SL-INFINITY Fans
• Memory: Corsair Vengeance RGB Pro 128GB (4x32GB) ‎CMW128GX4M4E3200C16
• Video Card(s): Nvidia GeForce RTX 3080TI Founders Edition
• Storage: Samsung 980 Pro (main drive) and 970 Pro (storage drive) - Both 1TB each
• Display(s): MSI Creator PS321URV 32in 4K @60Hz - HDR600
• Case: Lian Li 011 Dynamic XL ROG Edition
• Audio Device(s): Creative Sound Blaster AE-9
• Power Supply: Seasonic Prime TX 1000
• Mouse: Logitech MX Master 3s
• Keyboard: Razer Huntsman Elite (Silent keys)

PS: Nvidia's arrogance ignores these issues as they have a false sense of superiority.

Want to make this post as maybe it will help other users that they think they got a faulty RTX 3000 cards, so if you have this similar issue you can breath relax because your graphic card its ok, the faulty are the nvidia drivers, like any version looks like only afect RTX 3000 series.

There is a issue with artifacts, black/white squares or dots flickering in desktop and 2D mode like more than 1 year old already with RTX 3000 series cards.

At first I tought my RTX 3070 was faulty, so I RMA and I receveid a new RTX 3070 but have same issue. I had the RTX 3070 with the issue from feb 2021 and now from july 2022 I have a new RTX 3080 and after 2 weeks of use, I got the issue again, also in the RTX 3080, so with this card, I tested 3 different cards that all of them has the issue. For me its triggered like 1 per month lets say, and you have to restart PC in order to fix it. I changed DP cable like 3 times aswell just in case.
There is another guy here that posted time ago this issue, he said he RMA his card aswell and the issue triggered for him aswell with the new graphic card.

The issue is very random, there is no chance to reproduce it, the funny fact is that it will have everytime same patron I mean everytime there are some squares that are flickering in desktop, if you move mouse they dissapear, apear again, flickering and most likely the nvidia driver will crash with display has stopped working and recovered. They thend to flicker at same spot every time. Here are some examples of the issue that I found from other users but all of them with RTX 3000 series cards.

1)

2)

3)

4)

5)

I tested bunch of different driver versions, like from beggining of 2021 that I got first time the issue, untill now so there are alot of drivers and all are triggering the issue in certain moment. I did fresh windows install like 10 times , tested windows 10, tested windows 11, nothing works.

Here is my issue https://www.techpowerup.com/forums/...er-in-desktop-and-2d-mode.297640/post-4923472

Try going into your BIOS and force PCIe Gen to Gen 3 (It's probably on Auto by default) and also test with an HDMI cable to see if it's just the display port

That doesn't work

I've seen this problem before with my 3090, but it only occurs with unstable memory settings. Very easy to reproduce for me, I just set memory to +1000 on Afterburner and try to play a game, after 20 minutes or so this happens without fail. How are your memory junction temperatures?

That is not the same issue as we have.

Can you try lowering your memory speed a little bit? Even if it's at 0, try like -200, see if it happens again. It's very weird, I know, games run fine, benchmarks score OK. But if my 3090's memory is beyond +800 (1319 MHz in GPU-Z) this artifacting on the desktop/2D always occurs. Of course, if it is happening stock, you have a problem somewhere. Maybe memory running hot, maybe your card needs an RMA.

This issue has nothing to do with any OC software.

After spending a year with my 3070Ti I can say I haven't seen this.

I think your OC is unstable..

If your card isn't overclocked, I would try as suggested.. lower your memory clock.

OC is not the issue as I have not OC'd my PC yet and I have not installed any OC software such as MSI Afterburner. Plus the 3070TI is not affected.

I’ve been using nvidia GPUs for a long time.. never saw that before unless GPU was pushed too hard.

I did see that rainbow road a couple of times in Forza Horizon 5 though lol..

3070TIs are not affected by this issue.

Then it's not the card.

You have a CryZen CPU and I don't think AMD has this issue.

 

[eidairaman1]

I have this exact same issue. For me, it comes intermittently when starting up my PC. It will show on the lock Screen, sign-in screen, and desktop. To get rid of it, all I need to do is reboot my system. An issue that also comes with this experience is that I cannot access the system BIOS when rebooting my PC where the monitor turns off before system BIOS POST has had a chance to complete. The monitor stays off until the lock Screen shows (The newest VBIOS that is supposed to fix a similar display issue DOES NOT WORK). Plus, the BIOS LED light stays white when it is supposed to be off. I have replaced this Video card twice and both have the same issues on multiple monitors and while using multiple DisplayPort cables. The latter issue doesn't happen when I turn on my PC or when the BIOS posts before the monitor turns off during a reboot.

See my system specs for details of what I have.

• System Name: Windows 11 Pro for Workstations Build
• Processor: Intel Core i9 9900K
• Motherboard: ASUS ROG MAXIMUS XI EXTREME
• Cooling: ASUS ROG STRIX LC II 360 ARGB AIO, 10x UNI FAN SL-INFINITY Fans
• Memory: Corsair Vengeance RGB Pro 128GB (4x32GB) ‎CMW128GX4M4E3200C16
• Video Card(s): Nvidia GeForce RTX 3080TI Founders Edition
• Storage: Samsung 980 Pro (main drive) and 970 Pro (storage drive) - Both 1TB each
• Display(s): MSI Creator PS321URV 32in 4K @60Hz - HDR600
• Case: Lian Li 011 Dynamic XL ROG Edition
• Audio Device(s): Creative Sound Blaster AE-9
• Power Supply: Seasonic Prime TX 1000
• Mouse: Logitech MX Master 3s
• Keyboard: Razer Huntsman Elite (Silent keys)

PS: Nvidia's arrogance ignores these issues as they have a false sense of superiority.


Here is my issue https://www.techpowerup.com/forums/...er-in-desktop-and-2d-mode.297640/post-4923472


That doesn't work


That is not the same issue as we have.


This issue has nothing to do with any OC software.


OC is not the issue as I have not OC'd my PC yet and I have not installed any OC software such as MSI Afterburner. Plus the 3070TI is not affected.


3070TIs are not affected by this issue.


You have a CryZen CPU and I don't think AMD has this issue.

Only thing i can suggest is a gpu-z screenshot from your card

 

[ThirtyNinety]

Only thing i can suggest is a gpu-z screenshot from your card

Really can't see how this will help him, as he has already replaced the card twice?
No, I rather would put the card in another pc, connected to the same peripherals, and see what happens next. As it seems the symptoms of his problem seem to be even more extensive than the ones I described here...

 

[BenchAndGames]

Really can't see how this will help him, as he has already replaced the card twice?
No, I rather would put the card in another pc, connected to the same peripherals, and see what happens next. As it seems the symptoms of his problem seem to be even more extensive than the ones I described here...

Its ok, the people that dont belive will never change his opinion, they will keep saying its a GPU failure, even if me, jesseinf and many others that replayed here,also on other posts are saying we already change the GPU with a new one, they will still saying its GPU failure. No point to talk with them ... better ignore

 

[ThirtyNinety]

Its ok, the people that dont belive will never change his opinion, they will keep saying its a GPU failure, even if me, jesseinf and many others that replayed here,also on other posts are saying we already change the GPU with a new one, they will still saying its GPU failure. No point to talk with them ... better ignore

Yeah it’s kinda sad, for me this is clearly a software-related one, specifically linked to what happens during Windows startup.

But I must say, it has been now one month since I saw the latest driver crash at Windows startup together with these small artifacts, whereas before I started to get the issue almost weekly.

 

[Conrado77]

It only happened to me once about a month ago and since then it hasn't happened again. The only change I have made is to disable fast startup. But with this I am not saying that this is the solution. Maybe it's just coincidence.

 

[ThirtyNinety]

It only happened to me once about a month ago and since then it hasn't happened again. The only change I have made is to disable fast startup. But with this I am not saying that this is the solution. Maybe it's just coincidence.

I believe so, as it was already disabled here when the issue started happening. I always disable this option as it is known to trigger gpu driver errors. But in the mean time still did not see the problem again since 6th of Dec.

 

[pavle]

This looks like some design peculiarity that nvidians haven't worked around yet. Interesting though.

 

[Vayra86]

I have no experience with the newer cards, but I did upgrade my driver since a long time lately for Darktide, and ever since, when I want to conjure up the NVCP, I can be left just waiting and it won't come up. The process starts, but won't show NVCP.

Only a restart of the PC fixes it. This does lend credence to the idea Nvidia's drivers can act differently on each boot.