I built a new PC with an Intel Ultra 7 270K Plus, an RTX 5080, 32GB of Corsair DDR5-6000 memory, and an ASUS Z890-P WiFi motherboard. I'm experiencing recurring blue screens and application crashes that appear to be related to memory instability.
MemTest86 produces the following results:
- With both modules installed in slots 2 and 4, Test 3 reaches 10,000 errors, all reported on bit 4000.
- Each module works without errors when tested by itself.
- Testing the modules in slots 1 and 3 together produces the same failures.
- After reseating the CPU, the errors and crashes disappear temporarily, but the memory fails again about three days later.
I have tested the individual modules in each slot, and the errors only occur when both modules are installed. The system has been doing this for more than a month, so I'm trying to determine whether the problem is defective RAM, the motherboard, the CPU's integrated memory controller, socket contact, or cooler pressure before replacing expensive parts.
4 Answers
Since each stick passes by itself but the system fails with two installed, the problem is more likely related to dual-channel operation than to a completely dead memory module. Start by loading BIOS defaults and disabling XMP or any memory overclock. Run both sticks at the board’s default JEDEC speed and retest. If that passes, the DDR5-6000 profile may simply be unstable with this CPU and motherboard combination, even if it is advertised as supported.
Update the motherboard BIOS to the newest stable version, clear CMOS afterward, and use conservative memory settings while troubleshooting. Test with a single known-good module in the recommended slot, then test both modules together. If both modules pass individually but fail together at stock settings after the socket and BIOS checks, the likely suspects are the motherboard or the CPU’s integrated memory controller. At that point, test with another compatible kit if possible or pursue warranty replacement rather than swapping several expensive parts at random.
The fact that reseating the CPU temporarily fixes it is a major clue. The memory-controller connections run through the CPU socket, so a slightly bent socket pin, contamination, uneven CPU seating, or excessive cooler pressure can cause problems only when both memory channels are active. Remove the cooler and CPU, inspect the socket carefully under bright light, check for bent pins or debris, then reinstall the CPU and cooler evenly without overtightening. Also confirm the motherboard is not flexing.
I would inspect the socket before buying replacement memory. A contact problem can disappear temporarily after disturbing the CPU and then return once the cooler pressure or board temperature changes.
Do not treat the Windows crash dumps as proof that Windows itself is at fault. Repeated MemTest86 errors mean the platform is corrupting data before the operating system can do anything about it. Leave the RTX 5080 and storage alone for now and concentrate on the RAM, motherboard slots, CPU socket, BIOS configuration, and CPU memory controller. Avoid using the machine for important work until a long memory test passes with both modules installed.
The failures also happen with the memory running at default settings, and they return a few days after reseating the CPU, so I’m concerned this may be more than an XMP issue.