I’ll start with the tldr; My son saved up money from his first job and just purchased himself a 6800XT Reference card a couple weeks ago. I cannot for the life of me get this thing to run games reliably. The only solution that I have found is downclocking the GPU to roughly 1800mhz using the Adrenaline app. Every stress test I can find seems to run fine at stock settings (or even OC’d using the auto OC), but multiple games crash more or less instantly. The confusing thing is all of the crashing issues go away if I downclock the card.
This PC is about 3 months old, it took a while to source a new GPU so I just had the old RX570 in it for about 3 months. He's had zero issues with with this PC during that time. Initially when he bought the 6800XT no other changes were made to the PC, I pulled the old card out, put the new one in, used DDU to install new drivers and from day 1 it was unreliable for many games but fine in every other scenario. I have spent evening after evening trying to find the problem and I’ve exhausted every solution I have come across on Reddit. I just want him to be able to use the card, he worked really hard for it.
Most of my testing has been using Rainbow 6 Siege so I’ll use it to refer to, but multiple games (not all) behave this way. When I try to load Rainbow 6 Siege, it will start to load, the loading spinner will spin for a bit, then freeze and the entire system becomes unresponsive for 8 seconds (tdrDelay is set to 8 sec), the screen goes black for a second and then Windows recovers.
The AMD Bug Report Tool will pop up and display the following error “AMD Software detected that a driver timeout has occurred on your system. An issue report has been created…..”
Event Viewer says "display driver amdwddmg stopped responding and has successfully recovered."
WhoCrashed lists the following error “A livedump triggered by dxgkrnl occurred. You may have problems with your graphics driver or hardware.”
The thing I notice is that if I run HWInfo with a 50ms update rate, and watch it as the game loads is that hangs the instant the GPU clock and vram clocks ramp up to full.
GPU temps hover in the high 60’s while gaming, but even running Furmark it runs stable at about 78 degree GPU temp, 86 junc, 92 hot spot, fan 1150 rpm
---
Games that crash at almost instantly include: Horizon Zero Dawn, Rainbow 6 Siege, Detroit: Become Human, MS Flight Sim.
Rainbow 6 Siege seems to crash at the exact instance it would finish loading and start rendering a 3d scene.
Horizon Zero Dawn will boot into the main menu but will crash as soon as the campaign loads the scene. Same with using the built in benchmark, it loads then crashes instantly.
Detroit: Become Human gets past the main menu but crashes about 1 second after getting into the game. It will render a handful of frames then hang.
Games that seem to run fine. Age of Empires 4, Battlefield V, even with full RT, Satisfactory.
Stress tests run fine.
---
These are the Solutions I have tried, I don't know what else to do.
I'm at a complete loss as to what else I can try.
- Windows 11 - fully up to date
- Ryzen 7 5800X - stock clocks
- Asus ROG Crosshair Hero VIII
- Be quiet! Dark Rock Pro cpu cooler
- 32gb 3600mhz CL 16 Crucial Ballistix Ram (4 x 8gb)
- Samsung 980 pro 500gb nvme (boot drive)
- 850 Evo 500 gb sata ssd and PNY 480GB sata ssd setup as single virtual drive
- Western Digital Red 12tb HD
- Corsair RM850x PSU (two 8 pin connectors, not using pigtail)
- Dell 3440x1440 60hrz monitor
This PC is about 3 months old, it took a while to source a new GPU so I just had the old RX570 in it for about 3 months. He's had zero issues with with this PC during that time. Initially when he bought the 6800XT no other changes were made to the PC, I pulled the old card out, put the new one in, used DDU to install new drivers and from day 1 it was unreliable for many games but fine in every other scenario. I have spent evening after evening trying to find the problem and I’ve exhausted every solution I have come across on Reddit. I just want him to be able to use the card, he worked really hard for it.
Most of my testing has been using Rainbow 6 Siege so I’ll use it to refer to, but multiple games (not all) behave this way. When I try to load Rainbow 6 Siege, it will start to load, the loading spinner will spin for a bit, then freeze and the entire system becomes unresponsive for 8 seconds (tdrDelay is set to 8 sec), the screen goes black for a second and then Windows recovers.
The AMD Bug Report Tool will pop up and display the following error “AMD Software detected that a driver timeout has occurred on your system. An issue report has been created…..”
Event Viewer says "display driver amdwddmg stopped responding and has successfully recovered."
WhoCrashed lists the following error “A livedump triggered by dxgkrnl occurred. You may have problems with your graphics driver or hardware.”
The thing I notice is that if I run HWInfo with a 50ms update rate, and watch it as the game loads is that hangs the instant the GPU clock and vram clocks ramp up to full.
GPU temps hover in the high 60’s while gaming, but even running Furmark it runs stable at about 78 degree GPU temp, 86 junc, 92 hot spot, fan 1150 rpm
---
Games that crash at almost instantly include: Horizon Zero Dawn, Rainbow 6 Siege, Detroit: Become Human, MS Flight Sim.
Rainbow 6 Siege seems to crash at the exact instance it would finish loading and start rendering a 3d scene.
Horizon Zero Dawn will boot into the main menu but will crash as soon as the campaign loads the scene. Same with using the built in benchmark, it loads then crashes instantly.
Detroit: Become Human gets past the main menu but crashes about 1 second after getting into the game. It will render a handful of frames then hang.
Games that seem to run fine. Age of Empires 4, Battlefield V, even with full RT, Satisfactory.
Stress tests run fine.
- Furmark - no issues
- Linpack - no errors
- Cinebench - runs with no issues
- Ran all the OCCT stress tests that were available (system is stable at stock speeds) - pass
---
These are the Solutions I have tried, I don't know what else to do.
- DDU to remove old drivers
- AMD Cleanup Utility
- Tried several AMD driver versions
- Installing driver versions with Adrenalin (including todays)
- Manually installing drivers via Device Manager
- Using the default Windows Update GPU drivers
- Uninstalled Ryzen Master
- Re-seated the GPU
- Changed GPU PCI slots
- Re-seated ram
- Various Ram configs (2 sticks, 4 stick, swapping sticks out)
- Turned off XMP
- Turned off SAM
- Turned off Fast Boot in bios
- Bought a new power supply (Went from Evga Supernova G3 650w to Corsair RM850x)
- Updated Chipset bios from 3904 to 4006 (4006 just came out)
- Registry entry to increase tdrDelay
- Reinstalled BattleEye
- Turned off the Ubisoft connect overlays
- Uninstalled MSI afterburner
- Ran DirectX Diagnostic Tool - no issues
- Windows memTest - passed
- Memtest86 - ran full 4 hour test suite in safe mode - passed
- Ran the windows system file checker “sfc /scannow” - no issues
- Disabled c-state
I'm at a complete loss as to what else I can try.