Weird CPU Problem (Need maximum processor power state at bellow 100% to work)

[Nertan Lucian]

Here are my minidumps: https://wetransfer.com/downloads/417965e3af66f8f663ff1bbed0704ef220180923115449/d633778f66e1e787f6a69f5a906d932720180923115449/79319a

CPU: AMD Phenom X6 1090T
GPU: ATI Radeo HD 6870
--

So I installed window 10 Pro. Needed to refresh my old windows
After install I get after a while BSOD (above in minidump) with various errors
I try different things from reinstalling my drivers, my windows, etc
Nothing helps, moreover analyzing mini dumps is completely out of my skills
Except I remember I had a similar BSOD problem a few years ago and the solution (random) was to throttle my CPU withing Power Settings, setting maximum CPU power state to 97%
To be sure, I set it to 95% . And it all works like a charm, no more BSOD, GPU works at full load perfectly, CPU can reach it's throttled max

It bugs me what is the issue? It seems I "lived" with this for years and it all went fine

Is it a CPU problem? If so, does someone know exactly why running at near 100% suddenly makes it to crash (makes no sense)
Is it a cooling issue? The CPU temperature seems to be fine after long hours of near full load (gaming, matlab, etc), it doesn't crash
Is it a PSU issue? Can my PSU (650W) be the issue being unable to supply enough wattage

I'm very curios and I though you guys

 

[Greeko85]

The only thing that comes to mind is that over time components such as motherboard VRMs may have degraded to a point that may have started to cause instability. More than likely I am wrong but hopefully one of the more experienced members can chime in.

 

[johnbl]

you have a service running that looks up a registry key then does a system call that is causing the bugcheck. unfortunately it is using a generic service name (svchost.exe) you will have to hunt through the services and find a non Microsoft service that is the cause and disable it.

you might also change the memory dump type to kernel and provide a kernel memory dump. it will have proper debugging info for the power management.

the system was up for 3 minutes.
you might check device manager for devices that were not configured.
(this info will also be in a kernel memory dump)
note: your bios is from 2012, this is in the date range of window 7. Windows 7 devices had power management functions disabled by default. these functions were turned on by default after windows 8 shipped. The problem is when they were turned on it expose all the bugs in the drivers and in the bios. New drivers will attempt to use these functions and may fail with a old bios version.

you should see if you can update the motherboard bios and drivers from the motherboard vendors website. If you can not update them you might want to disable the sleep functions. or remove the failing device service in the is case. (always better to fix the problem than remove the device/service)

setting the system to high performance mode will prevent power management functions from kicking in. I would look for a bios update and driver updates first.

 

[Nertan Lucian]

Ok.

So I switched to kernel dump. Hope I did this right.
Then I switched in power option for the maximum CPU power state to be 100%
After a few minutes I got the BSOD and here's the dump:

https://jumpshare.com/v/WHvwU7aWTLtPLCpJIBOu

--

So I know this is correlated with " maximum CPU power state to be 100% " because I run the PC since the initial post without bug at 95% and then, when I switched to 100%, a few minutes and was hit BSDO. So is i'm inclined to think is something related to hardware

 

[johnbl]

remove the overclock driver (below) it could be the source of the problem

----------
the file provided was a minidump. kernel dumps are stored in a different location and file name
c:\windows\memory.dmp
--------
chrome was running, it attempted to run something but it was not in memory, attempts to load it from the pagefile.sys resulted in a bugchek due to corruption. I can not tell the cause of the corruption.


you most likely should update to the current bios (from 2015)
it will have required updates for the 2012 usb 2.2 spec changes, the 2013 USB 3.0 spec changes and updates to the SATA changes. You have to do this since you are running the updated Microsoft generic drivers for these components. (or you will get strange problems)

you should also go into control panel and turn off the system virtual memory, reboot then go turn it back on. this will delete the current hidden c:\pagefile.sys. Sometimes these can be placed on bad sectors of the disk and result in corruption when the system attempts to copy files from memory to or from the disk.


you should also download and run crystaldiskinfo.exe to read the disk smart data from the drive. It will give you an idea if the drive is generating errors.

- generally if you get corruption you might run memtest86 from its own boot image to confirm your memory subsystem. I would update the BIOS first since it will reset your memory timings to the new defaults memory timings for your motherboard.

http://www.asrock.com/mb/AMD/960GM-VGS3%20FX/#Download

remove your overclocking software:
\C:\Program Files\AMD\ATI.ACE\Fuel\amd64\AODDriver2.sys Tue Feb 11 03:06:52 2014





notes:
very old driver:
Qualcomm Atheros AR8151 PCI-E Gigabit Ethernet Controller (NDIS 6.30)
\SystemRoot\System32\drivers\L1C63x64.sys Sun Mar 31 20:15:17 2013
this should be the update for this driver: (chipset = ar8151)
https://www.ath-drivers.eu/download-driver-for-Atheros-chipset-with-id-50-and-Windows10-64bit.html

machine info:
BIOS Version P1.20
BIOS Release Date 11/30/2012
Manufacturer ASRock
Product 960GM-VGS3 FX

Processor Version AMD Phenom(tm) II X6 1090T Processor
Processor Voltage 8eh - 1.4V
External Clock 200MHz
Max Speed 3200MHz
Current Speed 3200MHz

 

[Nertan Lucian]

The only thing that comes to mind is that over time components such as motherboard VRMs may have degraded to a point that may have started to cause instability. More than likely I am wrong but hopefully one of the more experienced members can chime in.

Did you heard of a similar problem when motherboards VRM were acting out? Is there a way to test if? If it's just the motherboard, I might consider getting another one