My custom-built PC has been freezing or restarting unexpectedly for nearly a year. It worked normally for about two months after I assembled it, then the problems began. Sometimes Windows reboots by itself; other times the display freezes and I have to hold the power button to shut the system down. During a freeze, the mouse may still move and audio from a video can continue briefly.
I initially suspected the XFX RX 580 8GB, so I tried different GPU drivers, including older versions. I also replaced the power supply, tested the memory with MemTest86, and monitored temperatures and voltages, but nothing obvious appeared. Running the system without the dedicated GPU has stopped the random restarts, although the same type of freeze still happens occasionally.
The issue is difficult to reproduce without the GPU. With the RX 580 installed, demanding games could sometimes trigger a restart within a few minutes, while the integrated graphics setup freezes at seemingly random times. Event Viewer only shows Kernel-Power 41, and stress tests have not consistently reproduced the problem.
The system uses a Ryzen 5 7600, XFX RX 580 8GB, 32GB of Kingston Fury Beast DDR5-5200 memory, an MSI B650M Gaming Plus WiFi motherboard, a 1TB Forgeon Nimbus Pro PCIe 4.0 NVMe SSD, and an MSI MAG A750GL 750W power supply. What should I test next to determine whether the issue is caused by the GPU, memory settings, motherboard, CPU, storage, or something else?
4 Answers
The fact that the problem was more repeatable during gaming points toward the RX 580, its driver, or its power connection, even though the freezes continue without it. Make sure the replacement PSU was installed with its own modular cables rather than reusing cables from the previous supply. Also check the GPU for physical or thermal issues and test with another known-good graphics card if possible.
Try returning the memory to completely standard JEDEC settings instead of A-XMP. DDR5-5200 may be an overclocked profile, and a memory test passing does not prove that the selected speed and timings are stable in everyday use. Leave it at stock settings for a week or two and see whether the freezes return. Updating the motherboard BIOS and AMD chipset drivers is also worthwhile.
Kernel-Power 41 usually only means the computer was not shut down cleanly; it does not identify the failed component. Check Reliability Monitor as well as Event Viewer, and look for graphics-driver, WHEA, storage, or other errors around the time of the freeze. Run separate stress tests after a cold boot, since the failure may not appear under every workload.
The RX 580 is an older design, so its age and condition could explain why gaming caused restarts. However, because the system still freezes with integrated graphics, do not assume the GPU is the only fault. Check for firmware updates for the NVMe drive, update the BIOS, test with one memory stick at a time, and consider temporarily booting from another drive if the issue persists.
I checked Event Viewer and only found Kernel-Power 41. Stress tests have not reliably triggered the problem, although demanding games made the restart easier to reproduce when the RX 580 was installed.