AMD hard computer crash occurred due to software issues or hardware faults.
AMD hard computer crash occurred due to software issues or hardware faults.
Hello, I've faced a peculiar issue for about two years now. My AM4 platform frequently crashes during gaming sessions. It never powers down during regular use; instead, the system shuts off completely while the RGB stays active. The only solution is a hard reset. I've experimented with various fixes—replacing RAM, GPU, power supply, updating BIOS, reinstalling the OS, checking DDU, repositioning the CPU, reseating M.2 drives, even removing Commander Pro to test for overvoltage on the 12V rail. Still nothing works. What's peculiar is how unpredictable this behavior is—sometimes it goes months without a crash, then suddenly every time I launch a game, it freezes after just a few minutes. After clearing memory and letting the system sit, the problems disappear temporarily. I've shared benchmark screenshots showing stable performance with no thermal or power concerns. The only noticeable overclock is on the GPU memory, which the 6900X handles well even at higher settings. I disabled the overclock for troubleshooting, but it still functioned normally. CPU overclock isn't active, RAM has XMP enabled. I play a mix of games—mainly casual titles like Destiny 2, League of Legends, American Truck Sim, Dead Cells—and generally nothing too intense. 2D games run smoothly without issues, while 3D games occasionally freeze (which led me to suspect GPU problems). After swapping in an older Radeon VII and my old RX 480, crashes persisted. Interestingly, I completed Cyberpunk without any interruptions. No error messages appear in Windows logs, no warnings in Corsairs GUI, and the CMOS battery seems to retain settings after shutdowns. I attached temperature and power readings for CPU, GPU, both showing full load during testing. Benchmarks lasted around an hour each. I previously thought my smart access memory might be the culprit, but it wasn't. Any advice would be greatly appreciated!
Have you looked at GPU temperatures while gaming? Your readings during Unigine sessions are normal, but some titles noticeably impact GPU temps, especially with extended playtime. The adrenaline mode often triggers a driver error, though your BIOS shouldn’t be reset unless there’s a RAM problem and even then it may fail to respond. I’m still trying to figure this out. In older BG3 versions, I experienced strange crashes every three minutes. With Unity games, the issue disappeared once I limited my GPU clock speed to 2400MHz (on a 7900XTX that boosts up to 2800/2900). You might want to try something similar.
Hello, the issue is likely around 5-10 degrees hotter or cooler based on the game. The maximum temperature usually reaches 85°C on the hotspot. That’s why I initially thought it was a GPU problem, since another GPU behaved similarly. After rebooting, I’d enter BIOS and see the default screen reset. In Windows, opening AMD Adrenaline would show a reset to defaults message. I tried adjusting the max frequency and it helped. BG3 experienced the same behavior.
85C is acceptable, but it should stop crashing before reaching 95C+ unless the issue is mainly thermal. Is this the highest temperature the GPU reaches? Are there any other areas with significant heat buildup, like VRAM?
I’d rather consider a RAM problem. GPU crashes don’t clear the BIOS, right? You can turn off XMP and see if it still causes issues. The issue seems to be with the RAM controller—my old Corsair Vengeance sticks on a similar board (Aorus X570 Pro) had major problems, even after replacing them, so I ended up sending a replacement kit.
I've experienced multiple hard resets of my 7800XT recently, each time requiring a GPU driver update. Since then, I always keep the drivers current and haven't faced any problems. Have you looked into your PSU as well to ensure it's functioning properly? Do you have a PSU tester or spare unit available for testing?
It's quite likely the issue is with the RAM controller on my board. I've tried a different GSKILL Triden 3200MHz setup and it still worked. My other PSU also succeeded. It might just need a manual adjustment to 3600 instead of using the XMP profile. I'll need to test to confirm the problem. It seems like it could linger for months before fully recovering. Right now it's not functioning properly—it's really strange.
All the issues you've checked have been resolved. It looks like a faulty CPU or motherboard might be the cause, since those components haven't been fully replaced for testing. Another thought was a failing SSD, which could lead to crashes if storage is unstable. Have you verified the smart status of your drives? Also, did you reinstall Windows on the same NvMe drive or experiment with different ones?