I have been dealing with this issue for several months mainly when playing prepar3d but its happened even when the computer is at idle. I have an extensive list of troubleshooted items and cannot seem to narrow it down.

To start here are my specs

•  RTX 2080ti 
• LG C9 55' 4K TV
• Overclocked 6700K to 4.7 ghz (no problems for years)
• 16 GB ram 
• 1200 W PSU
• Asus VIII ROG motherboard
• Win 10 fresh install 

Here is what I have troubleshooted

• Ran Furmark for over an hour on several occasions, no problems at all, max temp 76C with fans not even maxed out
• Uninstalled and re-installed Nvidia drivers multiple times (Nvidia support has worked with me for hours picking around at other drivers too)
• Fresh install of Win 10
• Unplugged all USB devices
• turned G-sync off 
• Used an HDMI cable (currently using display port) 

The part that really baffles me is I have an old 32" 1080p monitor that I have even got it to crash on.

I am getting the feeling that maybe its RAM related, I haven't properly tested the RAM.  Also, maybe its related to a setting on my 4K TV but like I said above I was able to duplicate on an old 1080p monitor. 

I am desperate to find the solution, any help would be greatly appreciated.  

Why don't you think it is the video card? IF you have problems on both monitors, I would suspect that.  I had a 1080ti go bad and it drove me crazy trying to troubleshoot. It was replaced under warranty and those problems disappeared.

RTX 2080 Titans, at least early models, had a problem related to its onboard Gddr6 memory chips. They were shipped with either Samsung or Micron chips. Cards with the Micron chips exhibited different issues, and yours seems similar to some. There are a few YouTube vids on the subject.

Thanks for the replies. I guess my only hesitance to 100% blame the GPU is that is runs Furmark and other stress testing programs fine for hours at good temps. My thought would be if there is any issues with the GPU, it would have a hard time doing this.

Thanks,

Dave

It doesn't appear that you have an actual GPU fan/heat issue. The error message is telling you that the "signal" is the problem, which suggests that either the OS or the graphics driver is losing communication with whatever circuitry monitors the "pulses" from the fan(s) that it uses to determine fan speed. I assume that "GPU fans max" in the thread title means that when this happens, the fans spin up to maximum speed, which seems consistent with my theory. If the card (or system) cannot monitor the fan speed, then the safest thing to do in all circumstances would be to spin them up to max RPMs. If the system can't throttle them properly, run them full blast just to make sure its not overheating.
 

It sounds like a hardware problem in the card to me. Too bad you don't have an identical card to swap it out with.

Edited July 25, 20205 yr by MDFlier

With my Phillips 4K UHD tv used as monitor if I do not use the shielded HDMI cable supplied, it can go off often

I went a good 4 days without a crash and was putting some good loads on the GPU with prepar3d. I finally found a scenario that can consistently recreate the problem. I'll try doing some more GPU stress tests with furmak. The really weird thing is that the problem comes in fits. It'll pass everything I throw at it for days and then it'll have a really bad day, crashing all the time even with light loads. I'm having some power issues with at least one USB port (wireless keyboard with fresh batteries missing keystrokes) Don't know if that somehow related, maybe a power issue. But CPU is overclocked nicely and stable, no blue screens so who knows? 

Thanks for the reply's guys,

Dave

The CPU controls the PCIE throughput if it start`s to develop a problem it can affect you GPU. 

Dave,

How big is your power supply? Maybe you're maxing it out?

12 minutes ago, MDFlier said:

Dave,

How big is your power supply? Maybe you're maxing it out?

A good point, but he wrote it in his first post: 1200 Watt.

Even its a crapy one he should have stabel enough power.

Yep. Missed that

Having the same problem, GTX 1070 8GB, 1200 watt corsair supply, 2 monitors, Win 7 pro.
Running anything or nothing at all.. Not real bad as it happens only 2 to 3 times per week with computer use @~7 hours per day.
It's like the card just turns off. -or-
Remember reading a while ago ( don't remember the specifics ) that the drivers have a built in failure recovery on driver error. It seems like it is trying to do this but doesn't fully finish.
Also, when this happens my system seems to continue to run.
This has been working for me -
When it happens, wait about 15 seconds, CTRL-ALT-DEL until a logon screen ( with the other options) shows, not a hard shutdown, log back on. Voila! Working again.
I am thinking the driver recovery is killing explorer.exe and the re-logon will get it back running again.
 

I have since tested my PSU and my memory, all testing well. I have also updated my motherboard's bios but got the crash even while the computer was at idle. I came across this post on reddit, it's my case to a "T". It seems reverting back to the very first Nvidia driver (for RTX at least) was the issue. I'm on 411.63 and so far so good just curious why more people aren't experiencing it.  

 

On 7/29/2020 at 11:20 AM, tyda0501 said:

I have since tested my PSU and my memory, all testing well. I have also updated my motherboard's bios but got the crash even while the computer was at idle. I came across this post on reddit, it's my case to a "T". It seems reverting back to the very first Nvidia driver (for RTX at least) was the issue. I'm on 411.63 and so far so good just curious why more people aren't experiencing it.  

 

I am having this exact issue with my ROG RTX2080Ti. Currently using the Latest Nvidia Driver and W10B2004. I am going to try dropping back to the 411.63 Drivers now, I'll let you know how it goes. 

 

UPDATE - It looks like the 411.63 drivers are too old for my current Windows version. I guess I need to go back down to B1904 which will require a clean install. Not sure if I have time to do that so going to have to hold off. There is a long thread on the MS support forum about this, 6 pages and no input from them yet... One  of the reports directly specifies P3D, but seems to be a larger NVIDIA/Windows issue for all #D applications. 

https://answers.microsoft.com/en-us/windows/forum/all/event-65-appmodel-runtime/a2bfa9b8-f270-49e1-a20b-7ae513421743