Question GTX/RTX GPU: DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

[rreluzco]

I have done extensive searches on google, chatGPT and the list of people this has happened (still happens) to is endless. There is very long list of "possible" reasons and an equally long list of alleged fixes. As long as that list is, it the exact same list with not much variation. Despite the exact same error, it happens in a lot of different video cards, CPU motherboard combos and all different things have helped some resolve it, and others (like mysef) where not a single one of the suggestions have done anything to help.

The rig:

• Ryzen 5 3600
• Asrock B450M / AC
• Corsair Vengeance RGB PRO 64GB (4 16GB sticks) 3200Mhz
• Team Group SSD 1 GB (The older ones, not NVME)
• GTX 1660 TI 6GB <-------My nephew used the same model (not the same card...) and played all these games without issue, Darktide and Atomic heart specifically
• Off brand PSU, 500W gold
• Stock AMD cooler.

Games:

• Stardew valley runs with no issues
• Deep Rock Galactic runs with no issues
• Warframe, runs amazing when it does, then crashes after about 10 minutes, DXGI
• Darktide, same....settings are on low, frames are 50-60, runs excellent, then crashes after about 15 minutes, DXGI
• Atomic heart, same as Darktide
• Cyberpunk, same as Darktide (again, the game scales itself down for the hardware and when it runs its excellent...till 15 minutes later)

I understand that it seems to be happening with the higher end, more demanding games. However these games are fairly well optimized to run on a variety of cards, old and new. All of these games scanned my system and said it meets the minimum requirements just set on low

Steps:

1. -Remove and reinstall the GPU drivers using DDU, making sure they are the most updated
2. -Install old drivers (About 7 different ones: 566.36, 572.70, 576.28, 576.40, 576.52, 576.88)
3. -Registry edits (as described everywhere as "the fix" lol)
4. -Undervolt the GPU -50 to -100 to -150 to -200, tried all of them
5. -Updated BIOS, reset to default and disabled XMP (running now at 2666Mhz lol)
6. -Lowered game settings
7. -Disabled Anti Aliasing in global settings
8. -turned off latency settings
9. -Reinstalled the games, verify integrity etc.
10. -Put in a new NVME SSD
11. -Replaced the PSU with a new 650W 80Gold
12. -Replaced the cooler with a Wraith RGB cooler (not GREAT per say but infinitely better than what was there)
13. -Tried it with GTX1080 (this one ran a tad hot, but never past its limit. It would hit 80-82 max and I did all the other stuff of undervolting etc) <---this particular card I played 100's of hours of Cyberpunk in a different system NEVER any issues
14. -Tried it with a brand new RTX 3060 (yep)
15. -Reinstalled windows

After all of this, the exact same, depressing, unwaverying crash, DXGI_ERROR_DEVICE_REMOVED: GetDeviceRemovedReason() = DXGI_ERROR_DEVICE_HUNG

Honestly I dont know what it is. All that is left to replace is the CPU and the motherboard and the memory. I have ordered a new Motherboard (B550 this time) and a new CPU (Ryzen 5 5600) but not the memory

I will post a result here if anyone is interested, I am beyond burnt

 

[Lutfij]

Team Group SSD 1 GB (The older ones, not NVME)
SATA based M.2 or 2.5" SSD?

Corsair Vengeance RGB PRO 64GB (4 16GB sticks) 3200Mhz
For the sake of relevance, if they are 4 sticks comprised of 2x 2x16GB kits, please look at their PCB revisions and report back.

Off brand PSU, 500W gold
This is probably the root of all your issues. Non branded, unreliably built PSU's are more of a dumpster fire than a PSU to power the entire system.

Remove and reinstall the GPU drivers using DDU, making sure they are the most updated
Users are advised to run DDU in Safe Mode to remove all GPU drivers(Intel, AMD and Nvidia) then manually install the latest driver(or a version of your choice) in an elevated command, i.e, Right click installer>Run as Administrator.

Updated BIOS, reset to default and disabled XMP (running now at 2666Mhz lol)
For the sake of relevance, please state the BIOS version for your motherboard at this moment of time.

Replaced the PSU with a new 650W 80Gold
650W and 80 Gold are the advertised wattage and 80+ efficiency ratings, respectively, what is the make and model of the PSU?

Reinstalled windows
Did you recreate your bootable USB installer to rule out a corruption? Installing the OS in offline mode, then installing all drivers in an elevated command while in offline mode?

I have ordered a new Motherboard (B550 this time) and a new CPU (Ryzen 5 5600) but not the memory
Speaking of memory, have you tried working with just two sticks of ram and seeing if you can reproduce the issue?

Lastly, have you tried taking your existing GPU and dropped into a donor machine to see if your GPU is indeed triggering the issue?

 

[rreluzco]

Team Group SSD 1 GB (The older ones, not NVME)
SATA based M.2 or 2.5" SSD?

Corsair Vengeance RGB PRO 64GB (4 16GB sticks) 3200Mhz
For the sake of relevance, if they are 4 sticks comprised of 2x 2x16GB kits, please look at their PCB revisions and report back.

Off brand PSU, 500W gold
This is probably the root of all your issues. Non branded, unreliably built PSU's are more of a dumpster fire than a PSU to power the entire system.

Remove and reinstall the GPU drivers using DDU, making sure they are the most updated
Users are advised to run DDU in Safe Mode to remove all GPU drivers(Intel, AMD and Nvidia) then manually install the latest driver(or a version of your choice) in an elevated command, i.e, Right click installer>Run as Administrator.

Updated BIOS, reset to default and disabled XMP (running now at 2666Mhz lol)
For the sake of relevance, please state the BIOS version for your motherboard at this moment of time.

Replaced the PSU with a new 650W 80Gold
650W and 80 Gold are the advertised wattage and 80+ efficiency ratings, respectively, what is the make and model of the PSU?

Reinstalled windows
Did you recreate your bootable USB installer to rule out a corruption? Installing the OS in offline mode, then installing all drivers in an elevated command while in offline mode?

I have ordered a new Motherboard (B550 this time) and a new CPU (Ryzen 5 5600) but not the memory
Speaking of memory, have you tried working with just two sticks of ram and seeing if you can reproduce the issue?

Lastly, have you tried taking your existing GPU and dropped into a donor machine to see if your GPU is indeed triggering the issue?

First, thank you so much for taking the time, greatly appreciated you have no idea.

SATA based M.2 or 2.5" SSD?
2.5" SSD, replaced it with a new, NVME of the same brand:
TEAMGROUP T-Force G50 1TB SLC Caching 3D TLC NAND NVMe InnoGrit PCIe Gen4x4 M.2 2280 Gaming SSD

What is the make and model of the PSU?
Segotep 650W 80 Plus Gold Certified Non-Modular ATX Power Supply

Users are advised to run DDU in Safe Mode to remove all GPU drivers........ Right click installer>Run as Administrator.
Yep, with 7 different drivers, new and old

For the sake of relevance, please state the BIOS version for your motherboard at this moment of time.
10.41 date 4/11/2025

Did you recreate your bootable USB installer to rule out a corruption? Installing the OS in offline mode, then installing all drivers in an elevated command while in offline mode?
Yes, A Bootable USB installer was created on a different machine, installed windows offline and had a separate USB with all of the motherboard drivers previously downloaded individually from ASROCK

Speaking of memory, have you tried working with just two sticks of ram and seeing if you can reproduce the issue?
I have not

Lastly, have you tried taking your existing GPU and dropped into a donor machine to see if your GPU is indeed triggering the issue?
I have not. However, I did use 3 different GPU's in this same rig, all three produced the same error, same amount of time, same games.
GTX 1660 TI (original)
GTX 1080
RTX 3060

At this point I am hard pressed to think it was any of the GPUS and inclined to think its somewhere between the PCIe slot, and or the CPU itself. What makes me wonder is the multitude of different GPU/CPU system combos this happens with over the years (as far back as 13+ years ago) to this very day and still no real, concise consistent reason OR solution (at least that I can find)

EDIT:
missed this one:
For the sake of relevance, if they are 4 sticks comprised of 2x 2x16GB kits, please look at their PCB revisions and report back.
Where would I find it? Yes, it is 2 kits of 16 and 16. They were bought at different times though, same exact model numbers, everything. I bought one kit, then a few months later bought the second kit, just clicked "purchase again" on Amazon

 

[Lutfij]

Where would I find it?
On the side of your ram sticks;
https://www.corsair.com/us/en/explo...NEsrbns6MgDGtfZmyyrZ0qnAC-dQ0KgvFjxgpGnfWntz5
Corsair is the only brand that sources ram IC"s from multiple vendors.

I bought one kit, then a few months later bought the second kit, just clicked "purchase again" on Amazon
I'm suspecting two different IC'd ram kits. Work with one kit, instead of mixing two kits of ram.

GTX 1660 TI (original)
GTX 1080
RTX 3060
The problem here is that each card has a higher TDP than the next. To complicate things, the RTX 3000 series was known to have high transient load spikes.

I suggested looking at a donor motherboard since buying parts for troubleshooting will only result in you having brand new parts for yesteryears tech, to build a new system. There are instances where a particular ASRock or Gigabyte motherboard was the culprit in the entire picture.

 

[rreluzco]

New motherboard was installed today and that was it. Bad motherboard all this time. Tried all three GPU's and all of them worked perfect, zero error. Who knew