Hard crash detected, need assistance!
Hard crash detected, need assistance!
Over the past few days my 7900XTX hard crashes during gaming. I see a black screen, fans spin at full speed, need to perform a hard reset, and after that GPU drivers stop working so I have to reinstall or repair them. I tried the usual method—just reinstalling—and used DDU, but the GPU stays hot (around 62°C to 80°C). It started about 3-4 days ago after updating the Adenalin drivers to version 25.9. Playing Warhammer 3, I rolled back to 25.8 for a couple of days, but today it happens within minutes in several games. After updating back to 25.9 again, the issue persists. I've adjusted some voltage settings and removed them, but nothing changes. I'm really confused and don't know what's wrong.
Can you provide a method to easily recreate the issue? This would help eliminate other system problems. If it seems GPU-related, consider these steps: 1) Lower power settings significantly. I understand temperatures are fine, but focus on power supply stability. 2) Attempt overvoltage testing. If instability occurs, higher voltage might help; alternatively, adjust clock speeds at the same voltage. I don’t have your current settings.
It seems like the issue might be related to the GPU driver. If reverting to previous drivers doesn't help, it could be hardware-related. Check the Event Viewer for any DISM/SFC, chkdisk, or memtest results. Try running a Furmark test to see if it crashes. If it does, see if a Linux live USB causes the same problem. You might also want to test the GPU in another tower—perhaps the power supply is failing. I've experienced similar crashes linked to PSU issues before.
I've encountered problems with AMD drivers recently. Even without any visible text, I experienced a blank blue screen. I considered reverting to an older driver, but Adrenalin 25.5.1 caused a full BSOD. Newer drivers didn't resolve the issue either. It seems the problem is likely related to a driver.
SFC shows no problems, all drives appear normal in CrystalDiskInfo. The problem seems to be with the GPU, not memory or power supply. There’s no blue screen or shutdown, and even with a 850W PSU the load stays under 500W. It might be better for the GPU to perform optimally at its default settings before adjusting voltage.
It makes sense we're compensating Nvidia for their powerful GPUs... But there should be a solution to fix this, DDU, driver removal tools, Windows fixes, Linux... SteamOS?
It's puzzling because it functioned properly three days back then it starts failing again after reinstalling the same drivers.