Windows not loading the driver for the 7900 XTX model
Windows not loading the driver for the 7900 XTX model
For the past two weeks, I've been running into near-constant issues with the drivers for my 7900 XTX. The card has been in use since January and has been rock solid till I attempted to update to the drivers released in late July. Since then, immediately after almost every boot I get a message from the AMD driver stating the display driver timed out, and then an additional message saying that Crash Defender has identified an issue with the driver and it's now operating in a safe mode, and I need to reboot to re-enable full functionality. 90% of the time, rebooting puts me right back into the same cycle. During this entire process, I have to use the iGPU on my CPU (7900X) to get any display output, as my monitor just displays a 'no signal' when plugged into 7900 XTX. However, occasionally the card works fine after a reboot - no errors, no crashes, scores just about what's expected on benchmarks, and works fine for gaming. Then on the next boot, it'll be non-functional again. This randomness of functional vs non-functional is what has me worrying it's a hardware issue. I've also occasionally had the VGA error light on my motherboard turn on. However, it doesn't really seem to be linked with any behavior inside Windows. Throughout this whole process, Device Manager generally reports the 7900 XTX as either disabled, non-functioning with error 31, or does not see it at all. Strangely, every time I've booted into Safe Mode, the 7900 XTX is detected in Device Manager with no errors. Sometimes, the screen will go black and the system will reboot. It looks like a Kernel Power error 41 in Event Viewer, but I can't seem to find much more information. I've had two BSODs - once when attempting to update Windows (got a "DRIVER_POWER_STATE_FAILURE") and once when rebooting after installing AMD drivers (got a "SYSTEM_THREAD_EXCEPTION_NOT_HANDLED"). Both seem to be related to drivers. Here's a rundown of what I've tried: Windows integrity - fully updated system, disabled automatic driver updates to prevent it from screwing with the AMD drivers, ran sfc /scannow (no issues found), ran DISM (no issues found). No overclocking or hardware monitoring tools are running. Drivers - have reinstalled 3 different versions of AMD drivers (June, July, and August releases). Have directly reinstalled the driver, just loading up the installer in Windows and running it. Have booted to safe mode, run DDU and disabled all networking, before rebooting and attempting to install AMD drivers. Have booted to safe mode, run AMD's cleanup tool and disabled all networking, before rebooting and attempting to install AMD drivers. System - I've updated my BIOS to the latest version and have disabled all overclocks. I've also updated my chipset drivers to the latest version. Hardware - the GPU is properly seated in the PCIE slot. I'm using two separate power cables, both of which are properly plugged into a 1000w PSU. My RAM is installed in the correct slots. While not the most complete system stability check, I've run Prime95 for a while without any errors or crashes. This is really getting on my nerves at this point and has taken up way too much time. I'd really appreciate anything else to try, as I'm at my wits end. I've also attached a few screenshots of the errors I've experienced:
Would you like to test the GPU on a different machine? Have you installed it again? Which power supply unit are you using?
A fresh driver update is out now—consider giving the latest one a shot: https://www.techpowerup.com/312655/amd-s...l-released
I'm not sure about another system, but I'm trying to borrow something temporarily to test it. I've repositioned the card again, just to be sure. I have a brand new Corsair RM1000X PSU from January. I tried installing drivers in several ways—standard Windows setup, after updating DDU, using AMD's cleanup tool, and even in safe mode—but nothing worked.
I’m still confused about what’s happening. I reinserted the card, verified everything was wired correctly, and cleaned it up. After that, the GPU works fine. I’ve rebooted a few times and it’s running normally, scoring around 24500 in Time Spy and about 16600 in Superposition 1080p extreme. It’s hard to say if it’ll maintain performance, but everything seems okay. I don’t understand why a second physical reinstall resolved the driver problems or why the issues started in the first place. When I installed it, it seemed to fit a bit more on the right side than the left, but it’s still fully seated in the PCIE slot—just slightly angled compared to the covers. Since it’s a Lian Li Lancool III, there should be enough room.
It’s wise to inspect the PCI-E X16 port for any wear or deformation. The graphics card appears to be drooping, which could strain the port and cause improper alignment. Gently remove the screws at the back of the I/O cover, lift the GPU with your hand for stability, and then secure it again by tightening the screws.
It shows some wear, but nothing clearly wrong with the slot. Each time I reinstalled the card, I made sure to slightly lift it for better contact before securing the IO plate screws. The slot seems fine, though—my next suspect would be the motherboard if the card works perfectly on another system. Checking event logs revealed a "Kernel PnP error 411," which led to a post about similar problems and eventually pointing to faulty RAM. RAM is now the top priority before I test it on another machine.
Run the RAM through Windows memory diagnostics and HCI MemTest. Consider trying a bootable Memtest86, though no signs of faulty RAM have appeared yet. The GPU has operated perfectly over the last day—unusual.