Random crashes of AMD drivers aren't linked to overheating problems.
Random crashes of AMD drivers aren't linked to overheating problems.
Hello everyone, I've been browsing the forums for quite some time now. After roughly a year of ongoing issues, I'm reaching out for assistance. My gaming PC has repeatedly crashed the AMD display driver on a daily basis, and it seems unrelated to overheating. Common signs include sudden drops in gameplay (like Hearthstone), where the audio slows to 1-5 FPS and the mouse freezes, followed by a crash that restarts the game. Sometimes the same problem recurs during the same session (lasting 10-30 minutes). Other times, I experience no crashes for several hours. A reboot usually helps, or if the driver fails after a short period, it may stabilize afterward. However, I can assure you the PC will crash at least twice a day. Recently, while testing PRIME95 and Furmark, I noticed a crash after closing applications after about 30 seconds. I've tried several fixes: using DDU to remove drivers, updating GPU and chipset firmware, applying BIOS updates, disabling fTPM in BIOS, turning off C-States on ROG Strix UEFI, and running MEMTEST. GPU temps stayed around 75°C, CPU temps maxed at 70°C. Disabling XMP didn't help. I also disabled Multi-Plane Overlay, ran a MEMTEST for three hours with no failures, and adjusted RAM settings. The CPU was set to 3600X, GPU to ASUS B450i, and so on. If anyone has another idea or can guide me further, I'd really appreciate it! I'm starting to limit my gaming to Hearthstone and might consider switching to Xbox if needed. Edited February 24, 2024 by Rysh
Reviewing the event log reveals several issues: a critical hardware fault has been detected. Details include a processor core error originating from a machine check exception. For the game, Hearthstone.exe (version 28.6.2.62469) encountered faults with UnityPlayer.dll (version 2021.3.25.61228). Specific timestamps and paths are provided for further analysis.
Thanks for the feedback. I've run tests with Prime95 and Furmark simultaneously, encountering only minor issues after the initial stress phase. After that, I experienced a random crash during Cinebench and OCCT runs. I'll start testing with Cinebench and OCCT next.
If these tests pass, I’ll move on to reinstalling Windows 10. I’m on Windows 11 now, and I’m not sure if the issues I’m facing are more severe there than on 10, but it’s been over a year—who knows? I might also consider taking apart the PC and removing the PCIe raiser card, though I doubt that’s the cause since the Vega 56 uses PCIe 3.0.
It appears that enabling fTPM is necessary for installing Windows 11. After disabling it in the BIOS, Windows prompted me to create a new pin code, and once that was done, everything returned to normal but crashes persisted. I’ve tried updating my post about this issue, which links to NVIDIA support details. Regarding the CMOS battery voltage, checking it is important because low voltage can affect BIOS functionality, potentially causing save issues. My thought is that if the battery isn’t holding a stable voltage, the system might not properly apply changes.
You're correct about Bios, but UEFI behaves differently. I've noticed crashes, BSODs, and drive issues (like M.2 NVMe) that disappeared after a battery swap; usually, the voltage was between 2.2V and 2.6V. In roughly ten cases, system time and BIOS settings didn't reset, and no one pointed to the CR2032 battery as the cause. Each time I changed the battery, it fixed everything. I refer to these problems as the "UEFI battery AM4 mystery."
I can't directly access external events or their details unless you provide the information here. Please share the relevant data or link, and I'll help extract the Details tab for you.