My PC crashes unpredictably and sometimes behaves strangely before failing with MEMORY_MANAGEMENT blue screens. I analyzed several dump files with WinDbg, and they showed multiple 0x1A subtypes, including 0x404, 0x411, 0xA, and possibly 0x4E. The failures occurred in different memory-manager routines rather than repeating in one obvious location.
The used components are a Corsair 1000X power supply, an AMD Ryzen 9 5900X, an ASUS ROG Crosshair VIII Hero motherboard, and 2×16 GB G.Skill Trident Z DDR4-3600 memory. I have four memory sticks available and have tested each one separately, but the crashes still happen. All of these parts came from the same previous owner, so I am unsure whether the problem is the CPU, motherboard, memory controller, or something else.
What tests should I run to narrow down the faulty component?
3 Answers
Start by clearing the CMOS or loading BIOS defaults, then disable XMP/DOCP and test the system at stock memory settings. A leftover overclock or unstable memory profile can cause changing 0x1A errors. Run a bootable tool such as MemTest86 for several complete passes, testing one stick at a time in the primary slot specified by the motherboard manual. Then move that same stick between slots. If one stick fails in multiple slots, the stick is suspect; if failures follow a slot, the motherboard or socket may be at fault.
The different memory-management subtypes suggest a problem somewhere in the memory path, but a crash dump usually cannot identify the exact component. If each module works alone at default settings but the system fails with both installed, the integrated memory controller, motherboard, or CPU socket becomes more likely. Reseat the CPU and inspect the socket pins carefully under good lighting. A BIOS update and testing with known-good memory can also help separate the motherboard from the CPU.
OCCT passing does not rule out faulty memory, especially if it was run with an aggressive profile or for a limited time. Use a bootable memory test instead and let it complete at least four passes. Test methodically: defaults first, one module at a time, then both modules together. The most reliable final confirmation is swapping in known-good parts one at a time, since no software test can always distinguish a marginal CPU memory controller from a defective board.
Related Questions
Lenovo Thinkpad Stuck In Update Loop Install FilterDriverU2_Reload