Question Manageable BSOD turned into constant BSOD after attempting fix - ntoskrnl.exe ?

[ok_blinky]

For maybe a year now I've been getting regular BSOD crashes with no obvious hardware/software changes before it started. Crashes predictably occurred shortly after I'd first start using my PC for the day, almost always a little bit into playing a game like Overwatch, Diablo, or Counter Strike. After the crash it would usually work fine for the rest of the day. Sometimes it would crash overnight and occasionally while doing random tasks like browsing the internet.

At first I searched for some fixes, but none of the solutions I found in various threads worked for me. It was manageable enough and too difficult to fix, so I just accepted it would happen and went on like that for a while. Fast forward to today when I had some extra time and decided to fix it while just making everything worse.

BSOD Stop Codes would vary a lot, here's the list:

• PFN_LIST_CORRUPT
• MEMORY_MANAGEMENT
• IRQL_NOT_LESS_OR_EQUAL
• CRYPTO_LIBRARY_INTERNAL_ERROR
• POOL_CORRUPTION_IN_FILE_AREA
• SYSTEM_SERVICE_EXCEPTION
• CRITICAL_PROCESS_DIED
• DRIVER_VERIFIER_DETECTED_VIOLATION (after using this software)
• SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (see update #1)

Steps I took in attempt to fix the problem today:

• Updated Drivers (GPU & Windows Update)
• Ran Driver Verifier which caused it to crash much more often
• BlueScreenView always identifies ntoskrnl.exe
• Ran DDU to completely remove graphics/monitor drivers and start from scratch
• Ran ChkDsk until no errors were found
• Ran Memory Diagnostic with no errors found
• Ran CrystalDiskInfo to identify an "unhealthy" HDD that was just used for file storage
• Removed that "unhealthy" HDD from my PC
• Reinstalled Windows via Reset This PC
• Updated motherboard BIOS and all motherboard drivers (audio, network, chipset) directly from mfg
• Checked all physical connections to mobo and reseated RAM + GPU

Unfortunately, my crashes are a lot worse after taking the above steps, notably after the fresh Windows install a few hours ago. It BSOD 3-4 times while trying to download and install Diablo 4 + Overwatch. The first time it crashed this way, BlueScreenView had a second red flag for tcpip.sys, which is new. I updated the network drivers from the mobo mfg after that. Other odd behavior since the reinstall is that Overwatch itself is crashing -- not causing BSOD, just the application closing with an error -- on launch and shortly into playing if it makes it through launch. Diablo causes BSOD shortly into playing but that's when it would typically happen. Haven't had much time to monitor it since, but it did crash an hour ago with just a YouTube video playing while I searched tech support forums. Perhaps notably it has not crashed at all in the past hour while making this post with no video or games playing.

The PC was a prebuild from NZXT. Only hardware change I've made is upgrading the RAM and replacing the CPU Cooler that failed 2 years ago, long before the BSODs. It's possibly worth noting that the CPU Cooler failed immediately after I upgraded the RAM, which led me to perform a ton of diagnostics on my RAM before figuring out that actually the CPU Cooler was the problem.

Here's the link to my System Information and all the minidump files since the fresh Windows install: https://drive.google.com/drive/folders/1oadDPBfPe-KZnAryEpwn4saXanBRu_aw

Any help would be greatly appreciated. I've read a lot of posts with similar issues but nothing I could use to fix my problems. Happy to provide any more info as well.

Update #1 - 1 hour later:
Haven't been able to play a game for more than 10 mins without BSOD since the fresh Windows install. Oddly enough, I just had a brand new Stop Code and system file identified in the BSOD - dxgmms2.sys

Seemed like a notable development so I added the minidump file to the google drive link above.

 

[Colif]

• PFN_LIST_CORRUPT
• MEMORY_MANAGEMENT
• IRQL_NOT_LESS_OR_EQUAL
• CRYPTO_LIBRARY_INTERNAL_ERROR
• POOL_CORRUPTION_IN_FILE_AREA
• SYSTEM_SERVICE_EXCEPTION
• CRITICAL_PROCESS_DIED
• DRIVER_VERIFIER_DETECTED_VIOLATION (after using this software)
• SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (see update #1)

too many odd BSOD. Windows memory checker not best tool really.

Try running memtest86 on each of your ram sticks, one stick at a time, up to 4 passes. Only error count you want is 0, any higher could be cause of the BSOD. Remove/replace ram sticks with errors. Memtest is created as a bootable USB so that you don’t need windows to run it

update 1: directx. 90% of time I see it crash cause is GPU drivers

CPU: 5600X
MB: B450 AORUS PRO WIFI-CF
64gb G Skill 4 x F4-3600C16-16GVKC
RTX 3090
1tb Intel ssd
1tb WD HDD
1tb Samsung hdd
System info can't see your Power Supply

1st 5 errors - report

Spoiler

File: 102223-9781-01.dmp (Oct 23 2023 - 11:47:33)
BugCheck: [SYSTEM_SERVICE_EXCEPTION (3B)]
Probably caused by: memory_corruption (Process: Agent.exe)
Uptime: 0 Day(s), 1 Hour(s), 00 Min(s), and 02 Sec(s)

File: 102223-11781-01.dmp (Oct 23 2023 - 15:40:54)
BugCheck: [MEMORY_MANAGEMENT (1A)]
Probably caused by: memory_corruption (Process: chrome.exe)
Uptime: 0 Day(s), 1 Hour(s), 49 Min(s), and 33 Sec(s)

File: 102223-11671-01.dmp (Oct 23 2023 - 15:54:17)
BugCheck: [SYSTEM_THREAD_EXCEPTION_NOT_HANDLED_M (1000007E)]
Probably caused by: memory_corruption (Process: System)
Uptime: 0 Day(s), 0 Hour(s), 06 Min(s), and 22 Sec(s)

File: 102223-10828-01.dmp (Oct 23 2023 - 12:02:53)
BugCheck: [PFN_LIST_CORRUPT (4E)]
Probably caused by: memory_corruption (Process: System)
Uptime: 0 Day(s), 0 Hour(s), 14 Min(s), and 52 Sec(s)

File: 102223-10296-01.dmp (Oct 23 2023 - 12:05:31)
BugCheck: [CRITICAL_PROCESS_DIED (EF)]
Probably caused by: memory_corruption (Process: AWCC.Service.exe)
Uptime: 0 Day(s), 0 Hour(s), 02 Min(s), and 13 Sec(s)
Process mentioned are normally the victim, not cause.

last 3 errors - report

Spoiler

File: 102223-14390-01.dmp (Oct 23 2023 - 15:47:29)
BugCheck: [MEMORY_MANAGEMENT (1A)]
Probably caused by: memory_corruption (Process: System)
Uptime: 0 Day(s), 0 Hour(s), 06 Min(s), and 11 Sec(s)

File: 102223-12031-01.dmp (Oct 23 2023 - 12:53:07)
BugCheck: [PFN_LIST_CORRUPT (4E)]
Probably caused by: memory_corruption (Process: chrome.exe)
Uptime: 0 Day(s), 0 Hour(s), 15 Min(s), and 50 Sec(s)

File: 102223-11921-01.dmp (Oct 23 2023 - 13:50:56)
BugCheck: [MEMORY_MANAGEMENT (1A)]
Probably caused by: memory_corruption (Process: chrome.exe)
Uptime: 0 Day(s), 0 Hour(s), 23 Min(s), and 51 Sec(s)

Updating BIOS could help, you 4 behind - https://www.gigabyte.com/Motherboard/B450-AORUS-PRO-WIFI-rev-1x/support#support-dl-bios

try updating chipset drivers - https://www.amd.com/en/support/chipsets/amd-socket-am4/b450


Until you check ram with memtest, it could all be caused by bad ram.

 

[SkyNetRising]

Steps I took in attempt to fix the problem today:

• Ran Driver Verifier which caused it to crash much more often

First - you have to disable driver verifier.
It will cause BSODs just by being enabled.

Only hardware change I've made is upgrading the RAM and replacing the CPU Cooler

If you installed additional ram modules, they may cause instabilities, when running together with old modules.

Please list full specs of your system.
Show a photo of your system with side panel removed, if possible.
(upload to imgur.com and post link)

 

[Colif]

it appears to be one set

16384MB 2133MHz Unknown F4-3600C16-16GVKC
16384MB 2133MHz Unknown F4-3600C16-16GVKC
16384MB 2133MHz Unknown F4-3600C16-16GVKC
16384MB 2133MHz Unknown F4-3600C16-16GVKC

I can find them sold separate and together with that code.

 

[ok_blinky]

Thank you for such a detailed reply @Colif !

I updated my chipset drivers with the link you provided.

A bit confused by your comment saying the BIOS is out of date. I updated it before the original post and it shows I'm on the newest version (F64e - 09/20/2023). Am I missing something there?

One thing that seems weird, my RAM is a different model than what my System Information and memtest are showing. The RAM I have is 4x sticks of G Skill model F4-3600C16D-32GVKC, which matches the sticker on the RAM and what I purchased. Like you pointed out, it's showing as model F4-3600C16-16GVKC, which I can't find on the mfg site. Not sure why or if that matters though.

As you advised, I ran several memtests today and am getting a ton of errors there. Here's the summary:

• All 4 RAM Sticks = Maxed out errors and failed
• Stick #1 = Hundreds of errors soon into first pass (I stopped the testing to start the next one before leaving the house)
• Stick #2 = Maxed out errors and failed
• Stick #3 = Passed with 0 errors after 4 passes (took about 4 hours)
• Stick #4 = Exactly the same errors as #1 & #2

I repeated the memtest on each one and had the same results. Seems like 3 out of 4 sticks of my RAM just went bad? Then I popped in my old RAM (4x 8gb, slower speed) and ran memtest on that without any errors after 1 pass. Booted PC up with my old RAM and has been working for several hours without any issues playing games, etc., Will see if any pesky BSOD occurs.

Any other thoughts? I really appreciate you helping identify this so quickly. Can't believe it's as simple as RAM gone faulty. I ran memtest on these sticks when I got them 2 years ago without errors. Fingers crossed it stays stable and I can just upgrade RAM again.

 

[Colif]

A bit confused by your comment saying the BIOS is out of date. I updated it before the original post and it shows I'm on the newest version (F64e - 09/20/2023). Am I missing something there?

i was going from dump files. If you updated since then, fine.

• All 4 RAM Sticks = Maxed out errors and failed
• Stick #1 = Hundreds of errors soon into first pass (I stopped the testing to start the next one before leaving the house)
• Stick #2 = Maxed out errors and failed
• Stick #3 = Passed with 0 errors after 4 passes (took about 4 hours)
• Stick #4 = Exactly the same errors as #1 & #2

can't say i seen 4 sticks die but memtest doesn't lie twice.

The RAM I have is 4x sticks of G Skill model F4-3600C16D-32GVKC, which matches the sticker on the RAM and what I purchased.

Did you replace ram yourself or did people who made PC do it?

that code is for a 32gb set of ram. You seem to have mixed two sets together

F4-3600C16D-32GVKC

F4 = DDR4
3600 is speed
C16 is latency
d is the revision of the ram type, they release a new one every year or so.
32gb is capacity of the set
GVKC appears to just be its color.

Info - Meaning of RAM Product Item Numbers/Codes

If you are looking at differences between RAM modules being sold or if you are just checking some RAM models of your motherboard's Memory QVL that have been tested, you may have come across with these hard-to-figure-out RAM Product Items/Codes such as: "HX430C15SB2/8" or "TED416G2133C1501" or...

forums.tomshardware.com

So your sticks both being 32GVKC & (according to bios) 16GVKC means its either 4 different sets or 2 32gb sets stuck together. And AMD systems really don't like mixing.

I would make sure the ram you replace it with is 1 set. its code should include the number 64.
Ram in sets has been tested to work with only those sticks in that set.
The more you add that weren't tested, the more chance of errors. It wouldn't cause memtest errors but can cause BSOD due to timing differences.

Like you pointed out, it's showing as model F4-3600C16-16GVKC, which I can't find on the mfg site. Not sure why or if that matters though.

its there, you need to look for its revision - https://www.gskill.com/product/165/184/1562831134/F4-3600C16D-16GVKC

the D after C16 is the revision number. you won't find the original anymore but it did exist - https://www.memorybenchmark.net/ram.php?ram=G+Skill+Intl+F4-3600C16-16GVKC+16GB&id=13985

If you have proof of purchase for all the ram, you can send it with a screenshot of error page, and ram, back to G SKill and they should replace ram. But as stated above, mixing two sets can cause BSOD as well.

I think G Skill just use 16gb sticks that have one code on them and then make up sets of those sticks by testing the sticks together before including them in the set. That is only explanation I can give for why stickers and BIOS disagree.

 

[ubuysa]

I think it's most unlikely that 3 out of 4 RAM sticks are bad, I would expect that the problem lies elsewhere. Did you use the same slot to test each RAM stick? That may be at fault. The RAM may not be compatible with the board - is the RAM on the QVL for the board? It may the board itself that's faulty. It may even be the PSU.

Try running on the one RAM stick that passed Memtest86 for a few days, see what happens.