My PC recently started crashing with FAULTY_HARDWARE_CORRUPTED_PAGE and SYSTEM_SERVICE_EXCEPTION while I was just browsing the web. MemTest86 found no RAM errors, I checked the SSD, reinstalled the graphics drivers, and reviewed the crash logs. After lowering my memory speed from DDR4-3200 to DDR4-2933, the crashes appear to have stopped. The system had been stable for about eight months, and I never manually changed voltages or overclocked the CPU. My kit is 32GB using two 16GB modules rated at 16-20-20-40 and 1.35V. Could this indicate a degrading CPU memory controller, a motherboard problem, or simply an unstable memory configuration?
3 Answers
FAULTY_HARDWARE_CORRUPTED_PAGE indicates a memory-related fault, but it does not necessarily mean the physical RAM is bad. Windows may detect the corruption when reading data back from the page file, so the SSD or storage path can also be involved. SYSTEM_SERVICE_EXCEPTION is broad and can be caused by drivers, storage, memory instability, or other hardware. The crash dumps are much more useful than the stop-code names alone, so inspect several dumps for the consistently flagged module or driver.
It is possible for an overclocked memory configuration to become less stable over time, but lowering the speed does not prove that the CPU's memory controller has degraded. AMD officially rates the Ryzen 7 5700X for DDR4-3200 with two modules; higher speeds are overclocking, even if many systems can run them. If you were using four modules, the official limits would be lower. Also, memory advertised as 3200 or 3600 is technically measured in MT/s rather than actual MHz.
The system uses two 16GB modules at 1.35V, and I never adjusted the voltage or CPU settings. It had been stable at 3200 for months, which is why the sudden change seemed strange.
A system can become unstable without the CPU itself being permanently damaged. BIOS settings, memory training, temperature changes, a marginal motherboard slot, power delivery, or a driver or Windows change can all affect stability. Keep the lower speed for now, then test methodically: run each memory module separately, try the other motherboard slots, load BIOS defaults, check storage health and the page file, and test at the rated voltage with conservative timings. If errors return even at standard settings, investigate the RAM, motherboard, and CPU individually.
I uploaded the dump files and had already rolled back the graphics driver after testing it. Since reducing the memory speed stopped the crashes, I am wondering whether the memory controller is still the most likely cause.