Octopus Overlords

RTX 4080
Windows 11
Intel i9-13900K 3.00 GHz
32GB RAM
THERMALTAKE TOUGHPOWER GF3 1200W 80+ GOLD Power Supply

About once a day the two computer screens will go black and the GPU fans will ramp up to full speed. This can happen during low usage times like browsing or while playing a GPU intensive game. It requires a hard power shutdown and restart.

Troubleshooting already completed:

• Update graphics drivers.
  Update Windows.
  Update Monitor, BIOS, and motherboard drivers.
  Use only one monitor (it happens with both monitors individually or if they are both plugged in).
  Use different DisplayPort cables.
  Use 3 different kinds of surge protectors.
  Monitor the temp, usage, and fan speeds with Armoury Crate and HWMonitor. Both idle and heavy usage shows everything within expected ranges.
  Reset the GPU and power cable.

Any other suggestions?

Faulty card?

If you have not yet done so, try performing a CLEAN install of your graphics card drivers, by following the instructions detailed below by Tom's Hardware. This uses the Wagnard tools Display Driver Uninstaller to thoroughly cleanse prior rogue registry entries and incompletely uninstalled display driver cruft:

How to do a CLEAN install of your graphics card drivers using the Wagnard tools Display Driver Uninstaller

Damn. Yet another amazingly useful (and timely for me) post by AB. Thanks!

I have a running list of apps and software and troubleshooting steps you have posted - super helpful.

Carpet_pissr wrote: ↑Sat Jun 24, 2023 12:07 pm Damn. Yet another amazingly useful (and timely for me) post by AB. Thanks!

I have a running list of apps and software and troubleshooting steps you have posted - super helpful.

Anonymous Bosch wrote: ↑Sat Jun 24, 2023 12:31 pm

Carpet_pissr wrote: ↑Sat Jun 24, 2023 12:07 pm Damn. Yet another amazingly useful (and timely for me) post by AB. Thanks!

I have a running list of apps and software and troubleshooting steps you have posted - super helpful.

Two grumpy old man complaints:
1. Program strongly suggests booting into safe mode first. And wow has Win 11 really made that difficult for some reason compared to previous versions. :/
2. Once I’m ready to run the program, it shows “Select device type” and you choose Nvidia or AMD. ??!! Super ambiguous. Device whose drivers I want uninstalled? Current device installed? Ugh. And yes I’m aware that it’s probably just me and the way my brain works.

Carpet_pissr wrote: ↑Sat Jun 24, 2023 1:48 pm

Anonymous Bosch wrote: ↑Sat Jun 24, 2023 12:31 pm

Carpet_pissr wrote: ↑Sat Jun 24, 2023 12:07 pm Damn. Yet another amazingly useful (and timely for me) post by AB. Thanks!

I have a running list of apps and software and troubleshooting steps you have posted - super helpful.

Two grumpy old man complaints:
1. Program strongly suggests booting into safe mode first. And wow has Win 11 really made that difficult for some reason compared to previous versions. :/

Indeed, that sort of bullshit is least partly why my preferred gaming OS remains Windows 10 LTSC. However, this should steer you in the right direction for more easily getting Windows 11 into safe mode:

How to start Windows 11 in Safe Mode (8 ways) - Digital Citizen

Pretty simple..Nvidia card or AMD card. Its a driver uninstaller so thats what it will do.

Daehawk wrote: ↑Sat Jun 24, 2023 3:18 pm Pretty simple..Nvidia card or AMD card. Its a driver uninstaller so thats what it will do.

Right, but did you read my post? WHICH device? The one whose drivers I want to uninstall (in this case Nvidia), or the EXISTING device in my system (AMD) (to "protect" the drives associated with that, and not delete).

Should say something like "Drivers to be uninstalled"

If someone asks me to select a device, I am going to tend to think they mean my existing device, not one that is no longer in my system. And I also get that the dev is likely not a native English speaker, so completely get it, but it's these little details that kill me sometimes (and usually are related in some way to word usage or grammar).

I 'think' AB would mean to delete both and install a new one ..AMD if thats what you have now.

Yup, DDU is basically asking which particular system of drivers you wish to uninstall.

So, if you only want to remove Nvidia drivers, that is the appropriate choice to make. Also, if you no longer have or use Nvidia hardware, I would recommend selecting all of the preferences under the 'Nvidia Specific Options' in the options menu of DDU (and uncheck all of the 'AMD Specific Options' to "protect" the drivers and settings associated with that hardware). Then select 'GPU' and 'Nvidia' under device type, and click the 'Clean and Restart' button. Keep in mind though, Windows may need to re-detect and install your current display device(s) after you restart (which may require a subsequent restart).

BTW, if you'd prefer to install AMD Radeon software as cleanly as possible without unnecessary boat, I'd suggest downloading and extracting the Radeon Software Slimmer along with the latest drivers from AMD. Then use DDU to also remove your current AMD drivers, this time selecting all of the 'AMD Specific Options' to remove the drivers and settings associated with your current hardware. Then select 'GPU' and 'AMD' under device type. After you 'Clean and Restart,' run the RadeonSoftwareSlimmer.exe as Admin, click 'Pre Install' to specify where you downloaded the latest AMD drivers, and then you'll be able to specify precisely what you'd like installed. Here's a Youtube walk-through that shows Radeon Software Slimmer in action, which also provides some useful settings and recommendations that minimise telemetry and other unnecessary bloat. It was uploaded by a Russian with English subtitles however, so caveat emptor if that is a concern for you.

Nice. Having not had an AMD card in...many moons, was not aware that it was bloaty. Is it very?

Carpet_pissr wrote: ↑Sat Jun 24, 2023 6:04 pm Nice. Having not had an AMD card in...many moons, was not aware that it was bloaty. Is it very?

Yup, there's plenty of unnecessary telemetry and bloat with AMD drivers, too.

If you use Nvidia hardware, then NVCleanstall is the best way to install Nvidia drivers sans unnecessary boat, as shown here:

Thank you for all the suggestions. I replaced the AC Power cord from the case to the wall and so far my system has been stable for 2 days. Keeping fingers crossed it is not a bad GPU.

my (initial) goto answer for anything like this now is 'check the PSU'

hitbyambulance wrote: ↑Sun Jun 25, 2023 3:47 pm my (initial) goto answer for anything like this now is 'check the PSU'

It wasn't the PSU. It was the power cord that plugs into the wall. I used the one from my previous PC and so far it is working great. Three days without a black screen where previously it was happening at least once or more a day.

i was speaking generally, but:

Moliere wrote: ↑Sun Jun 25, 2023 6:27 pm

hitbyambulance wrote: ↑Sun Jun 25, 2023 3:47 pm my (initial) goto answer for anything like this now is 'check the PSU'

It wasn't the PSU. It was the power cord that plugs into the wall.

..that plugs into the PSU - it's all part of the mains power delivery system. (though it is ...odd that just changing the cable appears to have fixed the issue...)

And it's back. After 6 days with no black screen outages I got one yesterday. Time to swap out the GPU or PSU? Love these mystery computer problems.

Moliere wrote: ↑Fri Jun 30, 2023 11:40 am And it's back. After 6 days with no black screen outages I got one yesterday. Time to swap out the GPU or PSU? Love these mystery computer problems.

Or both..AH HA! Didn't think of that one did you?

Moliere wrote: ↑Fri Jun 30, 2023 11:40 am And it's back. After 6 days with no black screen outages I got one yesterday. Time to swap out the GPU or PSU? Love these mystery computer problems.

GPU

Moliere wrote:

hitbyambulance wrote: ↑Sun Jun 25, 2023 3:47 pm my (initial) goto answer for anything like this now is 'check the PSU'

It wasn't the PSU. It was the power cord that plugs into the wall. I used the one from my previous PC and so far it is working great. Three days without a black screen where previously it was happening at least once or more a day.

Maybe use a good power strip or ups; your wall jack might need conditioning.

Sent from my SM-S908U1 using Tapatalk

Or something is wrong with the PSU and the jostling that came with the cord switch knocked it back into place (temporarily.)

hitbyambulance wrote: ↑Sun Jun 25, 2023 3:47 pm my (initial) goto answer for anything like this now is 'check the PSU'

as in, try a different one

Push on all the power cords into the socket on the video card. It is a known thing, that the cords may not be fully seated. Not saying this is your issue, but one simple thing to try.

Montag wrote: ↑Tue Jul 04, 2023 1:27 pm Push on all the power cords into the socket on the video card. It is a known thing, that the cords may not be fully seated. Not saying this is your issue, but one simple thing to try.

That's what worries me, the unknown factor. It hasn't happened again since June 30th when I reset the GPU power connector. And of course I received the replacement PSU on Friday. Now I am reluctant to make any changes if it is working again.

Figured I'd post my recent similar experience and what I found.
I purchased a prebuild system with the intention of upgrading it (to ease my full build pain). So, upon arrival I upgraded a 4080 to a 4090 as well as increased memory and drive space of the new machine.
Everything was going fine until I experienced the same thing as you. I was running Stable Diffusion XL and when the the system attempted the Highres Fix function the GPU went out (could still hear the sound from the system so not everything went blank) and the PSU fan speed went full. Based on what I was doing I figured I stepped over the limit for the PSU (850w Thermaltake GF3). I bought a new GF 3A 1200 and after it came in I installed it and all was well. Over the 2 days it took to arrive I throttled my GPU to try to avoid this using MSI Afterburner. This seemed to help (maybe - I also went light on my usage of heavy GPU functions as I did think it could be a problem with too little power).
Anyway, at home my son only had a 750w PSU and a 3070 TI on one of our home systems so I figured I'd give the 850w to him and throw the 4080 I had extra into his system. Upon doing that - same thing! The only similar thing is the PSU, though. My work system never had the 4080 and 850w unit running together (installed the 4090 right away). I have built many systems, and being that this is the first time it has happened to me, and twice now with that particular PSU unit. Thus, I assume it is the PSU.
The new 1050 PSU Thermaltake is coming today and I am fairly sure the problem will be remedied by changing out the 850. I don't think the problem lies in the 850w limit, though. My son was playing Terraria when it first happened to him - not really a system eating game. Cyberpunk 2077 was able to run for about an hour for him until this happened again, so if it was just some GPU power draw problem it would have happened straight off. I think it is just a faulty PSU. I'll report back when the new PSU is installed to update.

I can report that my problem has not occurred again since resetting the GPU card and replacing the power supply. Maybe it was as simple as resetting the GPU, something I had done previously to no effect, or maybe it was the new power supply.

nixvivivi wrote: ↑Tue Aug 29, 2023 11:45 am Figured I'd post my recent similar experience and what I found.

Welcome to the forums. If you're a bot, you're a very very good one.

Kraken wrote: ↑Tue Aug 29, 2023 9:15 pm

nixvivivi wrote: ↑Tue Aug 29, 2023 11:45 am Figured I'd post my recent similar experience and what I found.

Welcome to the forums. If you're a bot, you're a very very good one.

We've got one trying to run a forum game.
They've come a long way baby.

Two months later and it started happening again. This time I replaced the power cable connecting the GPU to the PSU. So far the problem has not returned.

Moliere wrote: ↑Sun Jun 25, 2023 6:27 pm

hitbyambulance wrote: ↑Sun Jun 25, 2023 3:47 pm my (initial) goto answer for anything like this now is 'check the PSU'

It wasn't the PSU. It was the power cord that plugs into the wall. I used the one from my previous PC and so far it is working great. Three days without a black screen where previously it was happening at least once or more a day.

or the cable connecting the PSU to the GPU... the mitigation steps have been all PSU-adjacent, so far.