My PC has started shutting down or restarting within a few minutes of launching almost any game. The system has an Intel Core i9-12900K, a Seasonic 750W power supply, an RTX 5060 Ti, and 32GB of Corsair DDR4 memory. The problem began after I enabled XMP with four RAM sticks, so I disabled XMP and am now running two 16GB sticks at 2133 MT/s. I also recently replaced a drive that failed. Initially, the issue seemed to appear after switching Red Dead Redemption 2 from DirectX 12 to Vulkan, but now other games cause the same behavior. Sometimes the computer freezes completely instead of restarting, and I have also received an "UNEXPECTED KERNEL MODE TRAP" blue screen. Windows Memory Diagnostic found no RAM problems. CPU temperatures stay below about 72°C and GPU temperatures below 65°C. I'm having trouble keeping the system running long enough to use tools such as SFC. What should I check next to identify whether this is a power, hardware, driver, BIOS, or Windows problem?
2 Answers
Disabling Fast Startup is worth trying, but it probably is not the main cause of a crash that happens while launching games. Fast Startup only affects how Windows shuts down and resumes; turning it off forces a more complete shutdown and can rule out a corrupted hibernation state. For the actual diagnosis, try a clean graphics-driver reinstall, remove any GPU or CPU overclocks, and test with BIOS defaults. If the machine still loses power under load, reseat the GPU and power connectors and test with another known-good power supply if possible. The temperatures you reported look normal, so sudden power delivery, driver, firmware, or memory-controller instability remains more suspicious than overheating.
Check the reliability history first: press Win+R, type perfmon /rel, and look at the critical errors around the exact times of the crashes. Event Viewer can also help, but Reliability Monitor is usually easier to read. Use HWiNFO while running a game to watch CPU package power, GPU power, voltages, and temperatures rather than only the basic CPU and GPU temperatures. Also make sure automatic restart is disabled so a blue-screen error stays visible. Since this started after XMP, leave XMP disabled, use only two sticks in the recommended motherboard slots, load BIOS defaults, and update the BIOS and chipset drivers. A memory test passing once does not completely rule out instability from the motherboard or CPU memory controller.
I opened Event Viewer, but I can’t really tell which entries matter. I installed HWiNFO too, although I’m not sure what values from the log I should be looking at.