Memory OC remains consistent, yet you encounter page faults when accessing non-paged memory during startup only.
Memory OC remains consistent, yet you encounter page faults when accessing non-paged memory during startup only.
I've enhanced my memory from 3600 CL14 to 3733 CL14, but I still face random BSODs when booting Windows only and occasional Ethernet stuttering after startup. Memtest5, Antessa Extreme config, MemTest86, and DRAM calculator all passed without issues. A Prime95 test ran for 1 hour and 30 minutes. Everything seems stable except for occasional loading delays and intermittent BSODs during Windows updates, which reset the system. I've also tried using the Windows installation media, but it didn't fix the problem. What might be causing these issues and how should I proceed? Your specs: 5900x RTX 3080, X MSI x570, unified F4-3600C14-8GTZNB, H150i Elite Capellix.
RAM OCing is extremely challenging and can lead to many problems. It's hard to truly verify stability. I tried OCCT and used a memory testing tool on Linux, but after 18 hours of testing I encountered errors and had to tweak settings multiple times. It's not a pleasant process. You might want to avoid overclocking or revert to default settings to see if issues disappear. Check the guide carefully and read it again for more details.
You're looking for ways to assess potential issues before they cause instability during startup. Consider testing your setup under various conditions or using diagnostic tools to identify risks early.
It functions properly at the standard RAM speed. Stability is achieved, though minor performance drops might occur if the RAM isn<|pad|>, which can happen during setup. Corruption is unlikely unless there’s a more serious issue. Overclocking RAM isn’t recommended, as problems can cause significant performance loss. Your concern about flock not matching 1:1 is valid, but I have less experience with Ryzen compared to Intel.
It functions properly with the default XMPP profile, but since you only used it for a short period last week, there might still be concerns. Could something known to cause instability affect it? There are options to raise the profile settings if needed to help stabilize the system.
Corruption might already be present... if RAM altered bits in crucial files and saved them to storage, the situation is already resolved. However, if you assume it functions normally with default RAM settings... I’d guess your RAM isn’t stable enough, or your BIOS can’t handle the speed. Numerous voltage adjustments could improve stability, but RAM overclocking is complex and requires extensive testing. If Intel were involved, I’d experiment with System Agent and IO voltages. Still, more isn’t always better—sometimes extra settings backfire. The DRAM voltage also plays a role, but you might be applying RAM configurations it can’t support. I’ve tried various DRAM voltages with a lot of experimentation, but my units refused to cooperate. Please review the GitHub documentation carefully and consider what voltage options are actually available. Above all, have you performed consistent benchmarks before and after the RAM overclock? Based on what I understand about AMD, achieving a perfect 1:1 match between RAM and clock speed is crucial. This process could end up wasting time with no real benefit, only leading to frustration and potential OS issues.
Page faults occur due to unstable memory modules. I’m struggling to grasp the voltage levels you’re discussing. Why are they so high? It might just be that I’m not fully grasping it. Voltages seem reasonable, but everything feels a bit tight. 133ns is fine, though some components seem unusually constrained. I’m worried they might overheat and cause errors. I haven’t seen temperatures above 40°C, but I’m trying to keep them under 35°C.
Are you running on four sticks with strict timing? I'm not an expert, but that kind of BSOD might point to a RAM problem. Try resetting your RAM and see if Windows stops the crash or errors with programs/drivers. If everything works at default settings, your RAM is likely the issue—possibly due to tight timings. Also, if the VDDG voltage is 1.075V, your VSoC should be about 50mV higher, like 1.125V, to accommodate voltage regulation.