When I play demanding games such as Fortnite, the game suddenly crashes and the entire PC powers off without showing a blue screen. Afterward, Windows gets stuck cycling through ASUS Automatic Repair and Diagnosing your PC for around 20 minutes. I usually have to hold the power button, switch off the PSU, unplug the computer, and wait a few minutes before it boots normally again. Event Viewer only reports Kernel-Power Event ID 41, indicating an unexpected loss of power. The computer works normally for homework and other light tasks. I have monitored the GPU, which reaches only about 60°C, but I have not checked the CPU temperature yet. The system uses an RTX 3090, Ryzen 9 5900X, and an 850W 80 Plus Platinum PSU. What should I check first?
3 Answers
Start by checking temperatures for both the CPU and GPU while gaming. A GPU at 60°C is fine, but the CPU could still be overheating. Use a hardware monitor to watch CPU package temperature, GPU temperature, hotspot temperature, and power usage under load. If temperatures are normal, the next major suspect is the PSU, especially because gaming can create short power spikes that light workloads never trigger.
Also test system stability with BIOS defaults. Temporarily disable XMP or DOCP for the RAM, and remove any CPU or GPU overclock or undervolt. Unstable memory settings can appear fine during normal use but fail during heavy gaming. Updating the motherboard BIOS and chipset drivers may help too. If the system becomes stable with XMP disabled, try a lower memory speed rather than immediately returning to the previous setting.
An 850W Platinum PSU sounds reasonable on paper for a 3090 and 5900X, but the exact model, age, cabling, and condition still matter. The RTX 3090 can produce large transient spikes, so a failing or marginal PSU may shut the system off instantly under load. Make sure the GPU uses separate PCIe power cables rather than one daisy-chained cable, and if possible test with a known-good, high-quality PSU. Kernel-Power 41 only confirms that Windows lost power; it does not identify the cause.
The 850W rating alone doesn’t rule the PSU out. A failing unit or poor cabling can still cause this even when the wattage looks sufficient.
I have only monitored the GPU so far, and it stayed around 60°C. I’ll check the CPU and hotspot temperatures as well.