BSOD While CPU Under Load

Hi All,

I've been bashing my head against the wall trying to figure out this intermittent BSOD issue I have. It seems to mostly happening while the CPU is under intense load, but isn't really consistent. For example I played Destiny 2 pretty heavily for a few days without issue at more or less max settings, but trying to run a video render or even just playing WoW Classic while having Amazon Prime Video open on the other screen causing the system to crash out.

Typical BSOD analytical tools like WhoCrashed and BlueScreen viewer aren't really pointing to anything other than ntoskrnl.exe and ntkrnlmp.exe as the culprits.

I've tried running MemTest86+, which returned no errors after 11 passes, Intel's Processor Diagnostic utility, which passed all tests without issue, and FurMark, which also passed with no issue.

Prime95 however will not run a single test without a failure or system crash. I've run Smallest, Small, Large and Blend tests and they all either BSOD or crash the application itself.

I'm thinking I might have a bad CPU, but really would like a fresh set of eyes on this before having to put the money down to replace it.

Here's a OneDrive link to the collected logs:
Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

Thanks for the assistance!

CorsairLink.exe is just a program for managing the H115i CPU cooler. I don't ever mess around with overclocking or adjusting memory voltage so all of that should be the default already. I'll try uninstalling and see if that helps along with updating that nvidia driver and see if that helps.

- - - Updated - - -

After updating NVIDIA drivers and uninstalling the Corsair software as suggested, Prime95 did run for a significantly longer time than previously, however in the end still caused the system to crash.

Will next try driver verifier and use PC as normal.

Here are the logs from the most recent crash:

Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

There may be another crash in there from tonight, I had removed my processor to get the warranty information from it, cleaned it up a bit and applied new thermal paste, and ran Prime95 to see if that helped before reading your initial reply, but no dice there.

Also, please read the documentation that comes with Prime95
More info here (scroll down the page to the DISCUSSION section): Prime95

So I've run Driver Verifier for 2 days now with regular usage and no BSOD issues.

Testing with Prime95 afterwards is still producing system crashes. I've tried to follow the instructions linked by jdc1, but I'm honestly having a hard time following what the author is saying about pretty much anything.

Here's what I've been able to document at least. I ran each of the tests about 3 times each.

Running a Blend test BSOD's after about 6 and a half minutes.
Running Large FFT produces an error in the application which stops the test, also after about 6 and a half minutes.
Running Small FFT crashes the Prime95 application after about 30 seconds.

Here are the V2 logs collected after the Prime95 tests and crashes:

Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.


If I'm understanding that link correctly, that seems to point to a CPU issue?

There are many indications, but I will ask. What temperature does the CPU reach during testing when a crash occurs?

Had another crash while playing a game and watching a video in the other screen, after doing so for most of today.

Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

Didn't catch the temperature for that crash, but ran the Prime95 tests again afterwards and experienced another crash running the Blended test. CPU temp hovered around 70 degrees for both the blended and Large FFT, peaked at 90 degrees for small FFT before the application crashed.

I swapped out the memory for the system after the initial crash today with brand new sticks and had no change in behavior so I'm leaning to believe that the RAM is not the issue.

1) Open administrative command prompt and type or copy and paste:
2) sfc /scannow
3) dism /online /cleanup-image /scanhealth
4) dism /online /cleanup-image /restorehealth
5) sfc /scannow
6) chkdsk /scan
7) wmic recoveros set autoreboot = false
8) wmic recoveros set DebugInfoType = 7
9) bcdedit /enum {badmemory}

10) When these have completed > right click on the top bar or title bar of the administrative command prompt box > left click on edit then select all > right click on the top bar again > left click on edit then copy > paste into the thread

11) Make sure that there is no over clocking while troubleshooting.

12) Test all drives while in safe mode: (report into the thread if there are any BSOD while in safe mode)
Boot into Safe Mode on Windows 10

13) Run HD Tune (free version) (all drives)
HD Tune website
Post images into the thread for results on these tabs:
a) Health
b) Benchmark
c) Full error scan


14) Run Sea Tools for Windows
long generic test
Post an image of the test result into the thread
SeaTools for Windows |
Seagate
How to use SeaTools for Windows | Seagate Support US

15) Turn off Windows fast startup:

Turn On or Off Fast Startup in Windows 10

Have you checked if the thermal paste has dried? Have you cleaned the computer of dust (in particular the CPU and its cooling)?

Instructions for Driver Verifier get sorta complicated due to the many options available.
In short, the intent is to stress drivers. If a driver is faulty, the system is supposed to give a BSOD (one that is Driver Verifier enabled). And, hopefully, it'll reveal which driver was faulty.
But, there's lot's of reasons for it not happening - so don't sweat it.

Have you read this post about Prime95: Prime95
Scroll down to the DISCUSSION section and start reading there.

Also try the OCCT tests mentioned there (it's free from here: OCBASE / OCCT )

In short, Prime 95 tests:
- RAM
- CPU cache
- memory controller (built into the CPU)

So, it's either your RAM or your CPU (other faults are less likely, but still possible - such as a motherboard problem).
Can you borrow some RAM to test with? You have G-Skill RAM - which usually comes with a lifetime warranty.
Maybe contact G-Skill for replacements?
As for the CPU - do you have another compatible CPU that you can test with?

You may be able to purchase replacement RAM and/or CPU from a shop that will let you return it (for your money back) if it's not needed. Check BEFORE purchasing though!

Good luck!