Various BSODs randomly happening

Runtly · 04 Jan 2022

Win 10 21H2 (OS Build 19044.1415)

I have been experiencing various BSODs for the past month+, and have tried many different things to isolate the problem, but have reached the limit of my experience. I suspect a hardware issue, but would appreciate if someone more qualified takes a look. Blue screen error codes I have received include:
KMODE_EXCEPTION_NOT_HANDLED
STORE_DATA_STRUCTURE_CORRUPTION
SYSTEM_SERVICE_EXCEPTION
DRIVER_IRQL_NOT_LESS_OR_EQUAL
MEMORY_MANAGEMENT
PAGE_FAULT_IN_NONPAGED_AREA
ATTEMPTED_EXECUTE_OF_NOEXECUTE_MEMORY
SYSTEM_THREAD_EXCEPTION_NOT_HANDLED
UNEXPECTED_KERNEL_MODE_TRAP
KERNEL_SECURITY_CHECK_FAILURE

I have attached the log files from the v2 collector script.

Since these began, I have reinstalled Windows, run the system file check, Windows memory check, and a couple other hints I found here and on the official windows website. A few errors were corrected, but the problems persist.

The errors mostly happen while gaming or streaming video, but they do occasionally happen just while normally browsing the web. I have noticed that my fans (CPU mostly, I believe) have been running much more often than they used to, which is why I suspect hardware.

I have not changed the hardware on my machine in the past year.

Thanks in advance for any help!

MrPepka · 05 Jan 2022

Update NVIDIA GPU driver - NVIDIA DRIVERS Sterownik GeForce Game Ready WHQL
Test GPU with FurMark - FurMark - GPU Stress Test

Runtly · 05 Jan 2022

Thanks for the reply.

I'm assuming the Polish drivers you linked to aren't special in any way and upgraded to the latest English nVidia drivers. I've run the GPU stress test for 1.5 hrs with no issues. I'll try again later, as I've noticed that crashes tend to happen more often the longer Windows has been running.

Ztruker · 05 Jan 2022

The first thing to do is remove all tweaking. Restore all voltages, processor and memory speeds to nominal and see how to system behaves. If things improve then you can start applying your tweaks a little at a time and see which one causes the problem.

Is the RAM is on the Qualified Vendor List (QVL). I can't find it or I'd check. Even if it's not on it that doesn't mean it won't work correctly, just something to look at.

How long have you had the computer?

Runtly · 05 Jan 2022

I've never done any overclocking or anything myself. It's possible there's some automatic systems somewhere given that it's Alienware, but I've no idea where that might be.

I've had this computer for just over 4 years now.

The RAM is 2x 16GB Samsung DDR4-2666 P/N M378A2K43CB1-CTD

Ztruker · 06 Jan 2022

Okay, since you have not made any changes and it's run okay for 4 years, try running Driver Verifier:

Run Driver Verifier

Driver Verifier-- tracking down a mis-behaving driver.

Following the instructions in the above link will check for problems in all non-Microsoft drivers.

What we're looking for is a verifier generated BSOD with a mini dump that will tell us what driver caused it. If you get a BSOD, rerun the V2 log collector as soon as possible and upload the resulting zip file. Also see if there is a new C:\Windows \MEMORY.DMP file. If there is, copy it to another location then zip it and upload to a file sharing site like OneDrive and post a link to it here.

To check if verifier is active, open a Command Prompt and enter:

verifier /query

If not active it will respond with

No drivers are currently verified.

Follow instructions here to decide which drivers you want to test. 3-5 is a good start. Add 3-5 more every two hours until all are selected or you get a BSOD.

Enable and Disable Driver Verifier in Windows 10

Familiarize yourself with how to get out of Driver verifier from the above link and from the first link.ier, see what that shows you;

Runtly · 06 Jan 2022

I ran the verifier for a while, and eventually got a "normal" BSOD of ATTEMPTED_EXECUTE_OF_NOEXECUTE_MEMORY in win32kbase.sys. Here's the dump file that was created. I've also run the collector again and here's the results: Attachment 356596

I'll run it again tomorrow to get a verifier induced BSOD if possible.

Ztruker · 07 Jan 2022

The dump showed a fail in dxgkrnl.sys which is part of the video system.

You have a NVIDIA GeForce GTX 1070 Ti and a Intel(R) UHD Graphics 630. Are both being used?

If you are using just the Nvidia GPU then disable the Intel GPU in BIOS/EFI Setup, see if that makes any difference.

Edit: There is also a newer Nvidia driver available for your GPU which you can try:

Version:	497.29 ^WHQL
Release Date:	2021.12.20

You currently have
Version: 497.09 ^WHQL
Release Date: Wed Dec 01, 2021

Runtly · 08 Jan 2022

I upgraded to the latest driver earlier, and my nvidia control panels shows 497.29, so I'm not sure why it still shows the earlier version. I'll try disabling the CPU graphics and see if that helps anything.

I didn't have time to stress test my drivers today.

edit: I don't see an option to disable it in the UEFI settings, but I can do it in the device manager, so I'm going with that for now.

- - - Updated - - -

In case it's useful, it crashed again this morning with a KMODE_EXCEPTION BSOD, and here is the dump file it generated.

Ztruker · 10 Jan 2022

Didn't get the dump file.