My PC occasionally shuts down and immediately restarts while gaming or running GPU-intensive applications. It happens when the GPU is near 99% usage, usually no more than once or twice per day. The problem started after I replaced my graphics card and power supply.
Current hardware:
- Gigabyte B760M DS3H AX DDR4 motherboard
- Intel Core i7-12700 processor
- NVIDIA GeForce RTX 5060 Ti graphics card — recently installed
- Corsair Vengeance LPX 32GB DDR4-3200 memory
- Samsung 980 Pro 1TB NVMe SSD
- Corsair CX650 power supply — recently installed; previously used a CX550
The UPS makes a clicking sound when the shutdown occurs, so I'm unsure whether the UPS is failing or having trouble handling the system's load. Could the issue be caused by the PSU, UPS, GPU temperatures, or sudden power spikes?
I've also had separate blue-screen problems when XMP is enabled, including BAD_POOL_CALLER and Kernel Mode Heap Corruption. The two memory modules have different revision numbers, so I'm considering replacing them with a matched kit. I'm especially concerned because unexpected shutdowns are causing unsaved work in After Effects.
5 Answers
The blue screens are worth treating as a separate stability problem. Disable XMP for now and test the system at default memory settings. Mixed RAM revisions or mismatched kits can cause memory corruption, and errors such as BAD_POOL_CALLER and Kernel Mode Heap Corruption can result from unstable RAM or its settings. Run a proper memory test and, if possible, use one matched memory kit rather than mixing modules.
A sudden restart under GPU load can also indicate a PSU protection trip caused by a transient power spike, a loose power connector, or a defective power supply. Reseat the GPU power cable and use a separate PSU cable where applicable. Monitor GPU and CPU temperatures while testing, and check Windows Event Viewer for Kernel-Power events. Those events confirm an unexpected power loss but usually won’t identify the exact cause. If it still happens when connected directly to the wall, testing with another reputable PSU would be the next step.
The UPS is a strong suspect, especially since it clicks at the exact moment the system loses power. Some UPS units cannot handle the brief load spikes produced by a modern GPU, even when the computer’s average wattage is within the UPS rating. Try running the PC temporarily from a properly grounded wall outlet or a basic surge-protected power strip. If the restarts stop, check the UPS’s watt/VA rating, battery condition, and waveform compatibility before replacing it.
For the blue screens, look in C:WindowsMinidump for crash dumps and analyze several of them rather than relying only on the error names. They may point to a driver or memory-related issue. Until the hardware is stable, disable XMP and avoid critical unsaved work during stress testing; enabling automatic project backups in After Effects would also reduce the risk of losing work.
Check whether the GPU is overheating or whether its driver installation became unstable after the upgrade. Use an overlay to log GPU temperature, hotspot temperature, clock speed, and power draw during a workload. A clean graphics-driver installation can also help. Since the shutdown is a complete power loss rather than a normal crash, prioritize the UPS, PSU, cables, and temperatures before focusing only on Windows.
I’ll test it without the UPS first and see whether the random restarts disappear.