How to deal with repeated BSOD related to nvlddmkm.sys
How to deal with repeated BSOD related to nvlddmkm.sys
Hello!
I've been dealing with this BSOD repeatedly over the past weeks or months, and I haven't managed to resolve it. The crashes appear unexpectedly, whether I'm idle, streaming on Twitch, browsing in Chrome, or opening any random image file. The signs are consistent: the mouse cursor freezes, background audio loops, and the system becomes unresponsive (sometimes with video glitches) before restarting by itself.
I ran WinDbg and it always points to "nvlddmkm.sys" as the issue. Here are my system details:
- Power Supply: Corsair RM850x
- Motherboard: ASUS ROG Strix Z490-E Gaming
- CPU: Intel i7-10700K
- RAM: Corsair Vengeance LPX 64GB (2x32GB) DDR4 3600MHz C18 (QVL verified)
- GPU: Gigabyte NVIDIA GeForce RTX 4060 GAMING OC (top PCIe x16 slot)
- SSD: Samsung 980 PRO NVMe M.2 SSD (bottom M.2_2 slot)
Some observations:
- BIOS is up to date (version 3001)
- GPU is brand new, about two months old
- I've had other BSODs before like "SYSTEM_THREAD_EXCEPTION_NOT_HANDLED", "DPC_WATCHDOG_VIOLATION", and "VIDEO_TDR_FAILURE". When using WinDbg, all dump files point to "nvlddmkm.sys".
I considered it might be a GPU or driver problem, especially since these crashes occurred with an older GPU and non-QVL RAM on Windows 10. After replacing the GPU with a new RTX 4060, using fresh QVL RAM, and wiping the drive, I installed a fresh Windows 11, but the "DRIVER_OVERRAN_STACK_BUFFER" error still appears occasionally.
I also thought about XMP 1 being the cause, which could trigger these crashes, so I disabled it. Since then, I haven't had any crashes for a week. However, I'm still curious—could this be a coincidence? Everything else seems normal, BIOS is default except for XMP.
Currently, I'm using XMP 2 for my RAM, which sets the voltage to 1.361 (instead of 1.35). So far, no crashes, but I'm still testing.
Could it be that enabling XMP 1 is actually responsible? But why would that matter?
Also, the system performance is generally good—I can run Prime95 + Furmark at the same time without any issues, and temperatures are normal.
Only this persistent "DRIVER_OVERRAN_STACK_BUFFER" crash remains...
I'm really stuck here. Any advice would be greatly appreciated.
simply swapping out the physical card won't attract new drivers. you'd need to manually delete or swap the current driver setup. execute DDU in Windows' Safe Mode and opt to remove all optional data. finally, reinstall the newest driver package directly from Nvidia, excluding GeForce Experience.
Hello and thank you for your response!
I understand, but I performed a complete reinstall of Windows to ensure there were no issues with drivers.
The most recent BSOD occurred a few days ago on a new Windows 11 system, and I allowed Windows to install the GPU driver (an older one from 2023). I’m now testing with the latest Nvidia driver, not using GeForce Experience.
Yes, I have already tried DDU multiple times in safe mode when I was running an older Windows 10 version.
Edit 1: There is still uncertainty about whether XMP is active in the room and if it might be connected to the previous problem.
Also, is it typical for the DRAM voltage to rise to 1.361V when XMP 2 is enabled, even though it’s set to 1.35V?
Edit 2: I checked other RAM settings in BIOS—when AUTO is selected (default), the RAM runs at 2133Mhz and the DRAM voltage fluctuates between 1.200V and 1.217V. Is this also normal?
the xmp profiles are set up and checked by the maker for those particular RAM sets.
i would stick with XMP unless you see other unusual problems.
if you're experiencing more heat than usual, or believe it's needed, just switch to manual settings and enter the RAM maker's recommended limits. but .011v seems unlikely to make a big impact.
many current motherboard makers have built-in profiles that usually increase voltages for their 'auto-overclock' or other 'performance boost' settings.
Thank you for your response. I should mention that I verified and XMP 1 also sets the DRAM voltage to 1.3500V while the BIOS reads 1.361V, so there is no discrepancy between XMP 1 and 2 in that aspect.
In conclusion, everything seems to function as intended.
Another question: wouldn't a non-working XMP profile, especially if it explicitly uses a QVL RAM kit, indicate a potential hardware problem?
You're encountering a BSOD with the message "VIDEO_SCHEDULER_INTERNAL_ERROR" linked to "nvlddmkm.sys" while simply browsing files in Windows Explorer. What steps can you take at this moment?
Hello again.
I disabled XMP a few days back and set the RAM to 2133 MHz, which isn’t perfect but seems fine.
Since then I haven’t experienced any crashes and the system has stayed stable for three days now. I’m pretty sure the XMP profiles might be causing the Nvidia driver to crash.
What could be the cause and what should I do next?
Also, could this be related to a motherboard issue or should I try another RAM kit?
Another thought is that when I updated my BIOS to version 3001 a couple of weeks ago, I didn’t clear CMOS. Should I do that?
Appreciate any advice!
Video drivers seem very responsive to even minor RAM fluctuations. This was true for me on team red, and similar to what you've experienced. Turning off XMP resolved the problem instantly, without needing a strong XMP setting.
Another update: The system has remained stable and BSOD-free for more than a week, with XMP still disabled. I am confident now that XMP was responsible for the crashing video driver. I have ordered another RAM kit that can reach 3200 mhz. Additional testing will be carried out, and I will provide an update later.