My PC sometimes shuts down and immediately restarts, occasionally with a brief blue screen, when loading into Total War: Warhammer 2 or 3, ending a turn in Civilization VI, or launching Cyberpunk 2077 just before the main menu. Less demanding games such as Deep Rock Galactic, DRG Survivor, and League of Legends run normally. CPU and RAM stress tests in OCCT have also run for about an hour at full load without crashing.
The system has an eight-year-old Intel i7-8700K, Gigabyte Z370 AORUS Gaming 5-CF motherboard, an RTX 3060 that is about three years old, 32GB of RAM that is about three years old, and an eight-year-old 750W Corsair power supply.
I have already upgraded to Windows 11, updated the BIOS and chipset drivers, and completely reinstalled the affected games. Event Viewer has not been very helpful, and I am not very experienced with troubleshooting hardware. Friends have suggested that the CPU or motherboard could be failing, although I am unsure how to test those parts. I also have not received any new minidump files since February. What would be the best way to identify the faulty component?
3 Answers
RAM is another reasonable possibility, especially since the memory is newer than the rest of the platform. Run MemTest86 from a bootable USB and let it complete multiple passes. Any errors usually point to a faulty module, unstable settings, or a motherboard slot. Test sticks individually if errors appear, and temporarily disable XMP while troubleshooting.
A sudden restart with little or no useful crash information often indicates a power or hardware-level failure rather than a normal game crash. Check whether Windows is configured to create small memory dumps and whether automatic restart is hiding the blue-screen message. Also look for a critical Kernel-Power event, though that only confirms the unexpected shutdown and does not identify the failed part. Since the problem happens during demanding game transitions, test the PSU and RAM first, then check GPU temperatures, remove any overclocks or undervolts, and test the graphics card separately if needed.
The power supply is a strong suspect. An aging PSU can become unstable when a game causes a sudden combined CPU and GPU power spike, even if a CPU-only stress test passes and the rated wattage looks sufficient. If possible, temporarily test with a known-good PSU and see whether the same games remain stable. Avoid using questionable modular cables from another PSU.
I think I have a spare PSU, so I’ll test that after finishing a memory test.
I’m running MemTest86 now. Hopefully it comes back clean, since replacing the RAM would be inconvenient, but it should be straightforward to test.