Possible defect detected or MSI B550-A Pro available.
Possible defect detected or MSI B550-A Pro available.
I recently got a 5800x with an MSI B550-A Pro. I keep experiencing BSODs due to "Memory management" errors. Here are my system details and the steps I've taken. I visited Canada Computers to replace my CPU, but they said it was fine. However, since I bought it with a bundle, there were issues during the swap. Could you suggest any other solutions to stabilize my system? I performed a clean Windows 10 installation twice using two separate drives. The setup included: CPU - AMD 5800x (no OC, default), Cooler - H115i Pro, Motherboard - MSI B550-A Pro, RAM - G.Skill Ripjaws 4 DDR4 (2400MHz CL15-15-15-35, 32GB at 2133MHz), F4 - 2 4 0 0 C 1 5 Q - 3 2 G, GPU - EVGA 3080 Ti (no OC), PSU - AX860i, SSDs - 850 EVO 500GB, RAID drives - WD Black 2TB RAID 0, WD Blue 1TB. I also installed the latest drivers, ran a SFC/SCAN, tested RAM with MemTest, updated BIOS, rolled back Nvidia drivers, changed GPUs, checked all drivers, updated Windows updates, installed Corsair drivers for the AX860i, switched cables, and ran chkdsk. I also added a RAM stick and verified it. Have you tried any other fixes?
Consider boosting the CPU voltage by raising SOC to 1.08v. Adjust timings in the menu, turn off "power down mode." Verify the SPD settings in BIOS and ensure DOCP (XMP) timings align with the table. Experiment with 1T or 2T command rate timing—some cards perform better at 2T. Slightly increase memory voltage to 1.25v to avoid voltage drops. If temperatures rise, disable CPB or PBO and test if it stabilizes the CPU until you upgrade to a cooler suited for the high-performance CPU. Good luck!
I adjusted the SOC to 1.09, observed HWM readings between 1.088 and 1.100, stabilizing around 1.090 (1.08 triggers 1.075). Switching between 1T and 2T had no effect, while the DRAM was configured at 1.260 V (1.25 raises it to 1.245). I haven’t found a DOCP setting yet; XMP remains disabled. The RAM is currently running at 2133MHz. Temperatures reached 83°C during Blender rendering, which is acceptable temporarily, though OC room remains minimal. Still experiencing BOSD with memory management issues. My testing routine involves one Blender render, one or two benchmarks (or RS6 Siege), a Division 2 benchmark at the slowest point it crashes, then a steep benchmark run. It either triggers a BSOD or stops immediately after quitting the game. This testing approach is the fastest I can achieve before a crash, which usually occurs in 45 minutes to 1.5 hours—sometimes faster if I play a game or browse the web—and it happens randomly.
Consider using 16-16-16-38 at 1T with 1.35v. If you opt for 2T, ensure gear down mode is off, as that’s the only option to maintain flat 15s without stalling.
It worked out, my motherboard accepted the new RAM kit I installed instead of the one it originally had.