Persistent issue with crashing games remains unresolved.
Persistent issue with crashing games remains unresolved.
I need help finding a fix for my PC that keeps crashing when playing games. I've been trying to figure out what's wrong for a while now. Every time I launch a demanding game, it stops working and sometimes shows an error about the graphics card. Path of Exile 2 freezes everything completely, the Last of Us crashes to desktop, Baldur's Gate 3 does the same, and even games like League of Legends and Warhammer 40k end with a crash. The Backrooms game also crashes, but not with an error. It's been happening since the 17th of December. I've tested many titles and only Path of Exile and League of Legends consistently crash. Mostly it works when I disable XMP profiles, use default BIOS settings, and run at lower clock speeds. But it's still rare. I've adjusted power settings, changed BIOS, updated firmware, and even tried different GPUs. Still no consistent crashes except during gaming. I've also checked temperatures, memory tests, and hardware specs, but nothing seems to be the main cause. I'm worried about buying a new card later, especially since my current setup is unstable. Could you help identify the root issue? Any tips would be really appreciated.
Use one stick or RAM piece at a time and test various slots. It might be your RAM that’s not working. Check where your PSU is connected and consider using a surge protector. Switch out of the room if possible to see if it helps. Clear all undervolts and underpower settings, then run everything with default values—no XMP or extra tweaks. Reinstall chipset drivers, AGESA (or equivalent) from the motherboard site, IO drivers, etc. Verify your SSD status using HWINFO64 to see if it’s failing.
Try one stick at a time, switching to different slots if needed. Most users rely on Sensor Only mode for HWINFO; it seems the Drive Failure and Drive Warning sections aren’t visible in your screenshots.
I reviewed the log from HWINFO64 during testing of The Last of Us at maximum settings (starting at 60 fps, then crashing). It highlighted several key points around row 220, which marks the crash. The GPU was operating at 577% of its TDP, a value that seems unrealistic when considered properly. This suggests possible power delivery problems or faulty sensor readings. Spikes in GPU power near the crash may point to a failing VRM on the GPU or inconsistent power from the motherboard’s PCIe port. Memory usage stayed steady at row 220, but there was a minor decrease in available memory (551 MB) compared to earlier entries. The crash might be linked to memory instability or failure during rapid changes. The GPU was heavily busy at 2155 ms of utilization at this point. A mismatch between workload and power supply could lead to instability under stress. The effective GPU clock stayed at 0 MHz, possibly due to a monitoring glitch, indicating possible shutdown because of instability. Possible reasons include unstable power delivery, faulty PCIe slot, or GPU VRM issues. If the same pattern appears in other PCIe slots, the problem likely lies with the GPU itself. Suggested actions: Adjust GPU power limits to 80-90% using tools like MSI Afterburner, test the GPU on another system, and run memory-specific diagnostics such as OCCT. Check for voltage rail fluctuations on the 12V supply during load changes.
I’m still figuring this out. The best approach might be adding a new GPU and testing, or swapping it with another system and checking. If your CPU includes built-in graphics, remove the GPU and see if the games run smoothly. If problems persist, it could point to the CPU, RAM, or motherboard. Since you don’t have an Intel CPU, CPU wear shouldn’t be the issue. Some games reported “Out of VRAM” when the GPU was removed, even though the hardware seemed fine.
Make sure all cables are firmly connected in every slot on both the motherboard and GPU. You might need to reinsert them. There was a case where users faced issues with deep connections for the 4080 and 4090 GPUs—just ensure they’re seated properly.
Earlier, I encountered a situation with faulty power connectors on the 4080 and 4090. It turned out the problem stemmed from not inserting them deeply enough.
Check your power supply: verify it has the correct voltage (120/240V) and that the switch matches your region. It’s unlikely two PSUs are faulty at the same time, so suspect GPU or its connection. If you’ve used different PSUs, consider whether a bad cable was involved. Also, confirm the LED on the GPU’s power indicator is blinking—some manuals note this as a sign of poor power delivery.
Checked 3080 on my wife's machine too—it crashed. Tested 3060 on my own PC, it worked fine. Probably the GPU is failing, so I’ll lower the power and voltage a bit. No XMP settings appear to help stability right now. I’ll hold off until the new cards arrive next month.
That's unfortunate, but you're definitely aware of the problem.