[SOLVED] Constant crashes and BSOD

[seroba22]

Hello everyone!

So, if you like puzzles i'm certain you'll like this one...
Since 2 days or so, my PC started having problems like it used to have 2 or 3 months ago.
Games just randomly close like i killed the process in the task manager, when watching videos on YT and doing something else e.g. running a program / moving files to an USB the PC just BSOD. Also, just when these problems started my Browser (Opera/Chrome) tabs specifically kept crashing.



PC Specs:

OS: Win10 Pro

CPU: Alderlake i9-12900K

Mainboard: Gigabyte Z690 Aorus Elite DDR-4

PSU: 750 Watt be quiet! Straight Power 11 Modular 80+ Gold

GPU: Nvidia Geforce RTX 3080ti Msi Trio 12GB

RAM: 32GB GSkill Ripjaw DDR4 ; 2x16GB


For CPU-Z pictures =>

https://imgur.com/a/45UPuW5

View: https://imgur.com/a/45UPuW5


For Minidump files =>
https://1drv.ms/u/s!AgNhThPtLIrOgWTJyfun6f7SOy6t?e=Uq637Y

The 2 oldest dump files should be without XMP and the latest 2 with XMP.
The XMP profile seems to disable itself after a crash, probably because the setting isn't saved since i didn't restart right after enabling the profile.

I've had like a couple different BSOD so far,
KERNEL_SECURITY_FAILURE
MEMORY_MANAGEMENT
are the two i still remember.


Theres multiple warnings in the Eventviewer and i don't know if you really need to know what's written in there.
Memtest86 and the Windows Tool didn't find any problems with the RAM.
I got a couple warnings that something was wrong with my C: drive, tried multiple fixes and i don't think anything has changed.

I have updated the bios yesterday and all the other drivers are up to date aswell.
I've tried with and without XMP profile. After a clean Windows install + Disabled XMP profile, the system seemed to be stable for like a month or two, until i ran Path of Exile in combination with 2 Autohotkey scripts, which led to my PC to BSOD like 1 or 2 weeks ago. Not too sure if this game could be the cause of this though lol


Something fun that happened yesterday too:
I was watching a YT video and moving files to an USB stick to prepare for a clean Windows install. While doing so, right at the end when almost all files were transfered, my mouse + keyboard stopped working and the video and audio were slowing down. Just like when i'd watch at 20% speed. Man i really hope someone can help me i'm getting really frustrated and desperate at this point, since this PC wasn't cheap but i only run into problems with it.


Thank you in advance for your valuable time and help!
Kind regards from germany!

 

[johnbl]

Tested with pagefile disabled ->
WH3 just crashed when loading into a battle (Tried it 3 times to be sure). While restarting i had a BSOD? Didn't really notice it tbh, could be because i wasn't paying attention.
Tested with fixed size of the pagefile (Startsize 4963MB - Max was 9000MB) ->
I could load into battle but it resulted in a BSOD after a couple minutes.

I was in the BIOS and tried to see how i could fix the frequency of the CPU... I have no idea which settings to tweak, there's literally hundreds of things to change.

Intel diagnostics tool does not run, because the CPU fails at the Brand check.

So i'm only getting BSOD when playing demanding games so far.

Minidumps:
https://1drv.ms/u/s!Ant94T9rZaRfcLnBzQm2aiIvTKI?e=4f0idx

graphics subsystem was doing steam cleanup and used a bad memory address.
I would start by removing this tool:
C:\WINDOWS\System32\drivers\gdrv3.sys Fri Nov 5 01:35:11 2021

looks like it is a gigabyte graphics tool, it might conflict with the intel graphics accelerator driver windows installed.

use autoruns from here and unclick the driver to test if it is the cause of the bugcheck:Autoruns for Windows - Windows Sysinternals | Microsoft Docs
file loaded from this location:
C:\WINDOWS\System32\drivers\gdrv3.sys

crash was in cleanup code in the video memory manager, heap data was being cleaned up and one entry was corrupted. this could be a app bug or some issue with the graphics memory. Only issue I could see would be the gdrv3.sys messing something before the clean up was called. (just guessing)
update to gpu drivers would also be something to check.

 

[seroba22]

So i just did a clean install on this machine and re-installed every single driver.
I deactivated gdrv3.sys like you suggested with autorun but i ran into the same problem again.

I will probably just bring this machine to the PC Shop near me, i'm too frustrated to keep trying since all I can do is annoy you guys over and over with this.
That being said, I want to thank all that have tried their best to help me and sacrifice their free time to help a stranger over the internet.

Here is the last Mindump file for anyone who is curious about it:
https://1drv.ms/u/s!Ant94T9rZaRfcRQk-WHFWetMjqw?e=erAtwH

 

[johnbl]

last bugcheck was running a game, the video memory manager was doing some cleanup and there was a error. here is a description of the error:

In the process of decrementing the reference counts for an I/O space region, its accounting node could not be found. Typically this means the argument range was never locked or has already been unlocked. Parameter 2 is the base I/O frame. Parameter 3 is the number of pages in the region, and parameter 4 is the specific I/O frame whose node could not be found.

it could mean that the area was already free. or the address was wrong for some reason.

maybe you can isolate the problem to graphics. it looks like you have the cpu graphics and nvidia. maybe pull your nvidia card and run on the cpu graphics (if the game will run) and see if you still bugcheck.
(or try a different gpu if you have access to one)

also make sure you have updated your game.

i would also make sure you have applied all of the windows updates.
your build number looks like, 22000 (from last year)and build 22000.556 is from march.
just in case there was a bug in the graphics memory manager.
=============
edit: ok, it does look like you have updated:
the build on your dxgmms2.sys
File version: 10.0.22000.588
---------
I would also check bios settings, make sure there is no automatic overclock setting. make sure the PCI bus is set to 100Mhz and is not overclocked.
your cpu indicated it is running at ~MHz = REG_DWORD 3187
I generally suspect some setting is wrong when I see uneven clock rates. (it might be normal for these newer CPUs, i don't know)
i9-12900K
Intel64 Family 6 Model 151 Stepping 2

new processor with lots of things that could go wrong:
Enhanced Intel SpeedStep® Technology
Intel® Speed Shift Technology
Intel® Turbo Boost Max Technology 3.0
Intel® Turbo Boost Technology
Intel® Deep Learning Boost (Intel® DL Boost)
Intel® Gaussian & Neural Accelerator
(guess these can change the cpu clock rates)

overall, I am not seeing a lot of people getting this bugcheck. I think I would try the CPU graphics driver and see if you get the bugcheck in the video memory manager. That will tell you if it is the GPU or not. I guess I would also turn off the sound support for the gpu just in case it is something in the hd video sound driver. (disable all sound devices that do not have speakers attached to them) only install the basic gpu driver. maybe disable any unique functions for your monitor. Guess there was a bug in windows related to automatic screen brightness setting for HDR that messed up a bunch of editing software, it was fixed in a windows update. I guess i would disable special functions on the monitor also to run as a basic system just in case you are hitting some other fringe case that has a bug. Ie refresh rate settings
windows also started doing more checks on the graphics system to fight malware. This might expose more driver and game bugs.
that were not detected in older build versions of windows

 

[seroba22]

Okay, so after countless hours of pain we finally found out what the problem was.
As @johnbl has already pointed out, there were many things that could be wrong with a new processor like the one that i have.


I did the following things:
-> Clean install for Win11 including drivers, SSD etc.
-> Only installed benchmark programs like Prime95, Heaven etc.

After running 1 hour of the benchmark programs and not noticing anything unusual (Temps were fine and no crashes) i decided to try a game. Total War: Warhammer III for that matter, since this game seemed to be very efficient at making my PC go poof. After running the game for about 20 minutes, the usual problems occured. The game or any background programs would randomly close and refuse to open, the PC would BSOD etc.

After 20 minutes the CPU would reach about 85° C, which is pretty bad but not as hot as you'd need it to be for the PC to shut down or BSOD. But the PC always did.
The CPU cooler and paste was not the problem, but going into the BIOS and turning the Intel® Turbo Boost Max Technology 3.0 AND Intel® Turbo Boost Technology options off was literally like day and night. Now when the PC is under heavy load, it hardly goes above 55° C. After the bluescreens were fixed, the programs randomly closing was easily fixed by getting rid of one of the RAM sticks, even though they passed the memtest86. Is it rare for RAM to pass memtest even though it's faulty?


Also, thanks to everyone who was kind enough to help me troubleshoot this and sacrifice their valuable time.
Big thanks to @johnbl for his in-depth analysis!


TLDR;
-> Heat or the new CPU technology was the problem caused by
Intel® Turbo Boost Max Technology 3.0 and
Intel® Turbo Boost Technology
Just turn them off.