My Windows 11 PC has become nearly unusable after weeks of increasingly frequent blue screens and application crashes. Chrome and Opera may crash or turn white shortly after opening, the keyboard sometimes disconnects and reconnects, and the system can crash within minutes of booting or while sitting idle.
The system has an Intel Core i5-12600K, Radeon RX 7800 XT, ASRock B760M Pro RS WiFi motherboard, 32GB of G.Skill DDR5-6000 memory, and approximately 1TB of internal storage. XMP is disabled, and there are no CPU or GPU overclocks or undervolts.
The reported stop codes and modules have varied, including BAD_POOL_CALLER involving cbfilter24.sys, FLTMGR.SYS, and amdkmdag.sys; CRITICAL_PROCESS_DIED involving ntoskrnl.exe; SYSTEM_SERVICE_EXCEPTION involving dxgmms2.sys; SYSTEM_THREAD_EXCEPTION_NOT_HANDLED; DRIVER_IRQL_NOT_LESS_OR_EQUAL involving amdkmdag.sys and Netwtw14.sys; and a DRIVER_VERIFIER_DMA_VIOLATION after an external hard drive was connected.
I have run Windows Memory Diagnostic, CHKDSK, SFC, and DISM. CHKDSK found no bad sectors, but SFC repeatedly reports corruption it cannot repair, and DISM fails with error 0x800f0915. The Windows Update repair reinstall also fails with error 0x8007000d. I used DDU in Safe Mode and reinstalled the AMD graphics driver, disabled the Fortect filesystem filter after cbfilter24.sys appeared in a dump, disconnected the external hard drive, and confirmed Driver Verifier is disabled.
Disabling the graphics card did not completely stop the crashes, although amdkmdag.sys and dxgmms2.sys have appeared in several dumps. I am unsure whether the root cause is defective RAM, storage or Windows corruption, the graphics card or its driver, the motherboard or USB controller, another third-party driver, or several problems at once. What would you test next before replacing parts?
5 Answers
Because SFC cannot repair files and DISM also fails, I would back up important data and perform a genuinely clean Windows installation rather than relying on the in-place repair option. Install only the chipset, network, and graphics drivers needed to test the machine. If it still blue-screens on a clean installation with the external drive disconnected, that strongly shifts the investigation toward RAM, the SSD, GPU, motherboard, or power delivery.
Start by updating the motherboard BIOS and loading its defaults afterward. Also verify that the memory is actually running at safe JEDEC settings rather than an aggressive automatic profile. If the crashes continue, test each RAM stick individually and run a longer bootable test such as MemTest86; Windows Memory Diagnostic is useful but can miss intermittent faults.
Before replacing expensive parts, reduce the system to a minimal configuration: one RAM module, the boot drive, basic USB devices, and the graphics card only if the processor's integrated graphics can be used for testing. Disconnect extra USB devices and expansion hardware. Change one thing at a time and keep notes, since rapidly changing several components makes intermittent faults difficult to identify.
Check the internal drive's SMART data and health report with a tool such as CrystalDiskInfo, and review the system and disk event logs. A failing SSD or storage/controller problem can cause corrupted system files, failed repairs, CRITICAL_PROCESS_DIED, and seemingly unrelated driver crashes. Keep the external drive disconnected until the base system is stable.
The different modules being blamed do not necessarily mean every listed driver is defective. amdkmdag, dxgmms2, network drivers, and filesystem drivers can all be victims of bad memory, storage corruption, or another kernel-level problem. After checking hardware, remove or uninstall any remaining third-party filter, security, tuning, or monitoring software and use the original manufacturer drivers.
That makes sense. I will test the memory and drive first, then use a clean installation as a way to separate Windows corruption from a hardware problem.