I built this PC a little over a month ago, and it recently started restarting without warning while playing Assassin's Creed Mirage. The restart happens after roughly two hours, with no blue screen, and Windows starts normally afterward. Other games, including Overwatch, Transport Fever 2, and A Plague Tale: Innocence, have not caused the problem. If I quit Mirage before the two-hour point and use another game or application, the system also seems stable.
The system has an Intel Arc B580, Ryzen 5 7500F, Gigabyte B650M Eagle WiFi 6E motherboard, one 16 GB Patriot Viper Elite 5 DDR5-6000 CL30 module, Seasonic Focus GX-850 power supply, and Windows 11. Temperatures appear normal: the GPU stays below 60°C at around 70% usage, the CPU is slightly above 60°C at about 50% usage, and memory usage is around 11 GB.
I have updated the BIOS from F4 to F9, restored default BIOS settings, disabled EXPO, reinstalled the graphics driver with DDU, installed the AMD chipset drivers, and checked both SSDs. Event Viewer only shows Kernel-Power Event 41 with BugcheckCode 0, so there is no obvious crash dump or blue-screen error. What should I test next, and could this still indicate a hardware or power problem even though it only happens in one game?
2 Answers
The power supply is powerful enough on paper, but sudden resets with no bugcheck can still come from power delivery, cabling, or a PSU protection event. Confirm that the GPU uses the correct dedicated PCIe power cable and that all motherboard power connectors are fully seated; avoid adapters or loose extensions. Check the PSU's manufacturing date and, if the tests point in that direction, temporarily try a known-good modern PSU. Since the unit appears to be from around February 2026, age-related issues seem less likely, but a defective unit is still possible.
Kernel-Power 41 is usually just Windows reporting that the previous shutdown was unexpected; it does not identify the cause. Because there was no bugcheck and the code is zero, this looks more like a sudden power loss, hardware reset, or low-level hang than a normal driver crash. Check the surrounding System and Application events for WHEA, display-driver, or storage errors, and run repeatable stress tests for the CPU, GPU, memory, and combined system. A game can trigger a fault that other games never reach because the workload is different.
I tested both Balanced and High Performance power plans, but the restart happened with each. The event log still only shows Kernel-Power 41 and no WHEA errors.
The serial number begins with R2602, so I believe the PSU was made in February 2026. I will recheck the power cables and test the memory and GPU separately.