Windows 8 Random 0x0000001A errors with ntkrnlmp.exe

Code: 
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck 139, {3, ffffd00104d6a6e0, ffffd00104d6a638, 0}

*** WARNING: Unable to verify timestamp for bwcW8x64.sys
*** ERROR: Module load completed but symbols could not be loaded for bwcW8x64.sys
Probably caused by : bwcW8x64.sys ( bwcW8x64+4fda )

Followup: MachineOwner
Hi,
the above Bugcheck 139 means corruption has been detected in a critical data structure and two drivers were mentioned. As your aware most of the dumps point to this driver:
bwcW8x64.sys Fri Jan 22 22:07:20 2016: Killer Bandwidth Control Filter Driver please update:
Link Removed
or
Support For Z170A GAMING M3 | MSI Global | Motherboard - The world leader in motherboard design
I know from personal experience that when removing or installing a Killer driver then you need to disable any Anti-virus that you have running otherwise the driver may not install properly.

Please remove the current gpu driver using the display driver uninstaller. This will remove any corruption caused by former gpu driver installs and leave the system ready for a new gpu driver (or the same one again if it's the latest)

Open an admin command prompt and type:
sfc /scannow
Press enter and await results.

Post any new dump files.
 
Update:

I had my friend removal all Killer/Rivet software and manually install Just the network driver.

While using Driver Verifier he got a couple blue screens pointing to just nvlddmkm.sys, so we then ran DDU and then got the latest Nvidia drivers and installed only the graphics/HDAudio/PhysX drivers.

He made it through a game of Overwatch with no crashes, but it was largely unplayable due the jittering/clunky-ness from running driver verifier. Before uninstalling-reinstalling the display drivers he would get crashes/restarts in under 30 seconds.

Since windows didn't BSOD/Restart for the duration of a full game, we will have the computer run normally for a while and see what happens. Hopefully nothing out of the ordinary!
 
Nice job, it sounds like we may well be getting somewhere at last.. Please post any new dump files thank you.
 
Well that hope was short lived!

My friend played a game for an hour or so and then browsed the web for a bit before leaving to run some errands. He left the computer idle for a few hours and there was no BSODs during any of that activity.

However, when he returned he tried to fire up Overwatch and it crashed to windows within 15 seconds and moments later windows threw a BSOD. DMP file below.


I couldn't find any consistent information on dxgmms1.sys but it seems to be part of DirectX from what I can dig up. Since we ARE using the most recent drivers I find it difficult to believe they are the cause.

Since I was suspicious of the integrity of his system files, I had him run System File Checker(sfc) and the results read:
Windows Resource Protection found corrupt files but was unable to fix some of them. Details are included in the CBS.Log
Click to expand...

That can't be good! I have attached cbs.log as 7zip file because it's 9mb normally and these forums wont let me upload it as-is.

EDIT:
digging through the CBS file I found a couple points of interest;

2016-08-06 23:20:28, Info CSI 0000052e [SR] Could not reproject corrupted file [ml:520{260},l:102{51}]"\??\C:\Windows\Microsoft.NET\Framework64\v2.0.50727"\[l:34{17}]"System.Design.dll"; source file in store is also corrupted

2016-08-06 23:20:38, Info CSI 0000055f [SR] Could not reproject corrupted file [ml:520{260},l:128{64}]"\??\C:\Windows\Microsoft.NET\Framework64\v4.0.30319\NativeImages"\[l:36{18}]"system.core.ni.dll"; source file in store is also corrupted
 

Attachments

Last edited:
Code: 
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck 139, {3, ffffd001d921f3e0, ffffd001d921f338, 0}

Probably caused by : dxgmms1.sys ( dxgmms1!VIDMM_GLOBAL::CloseOneAllocation+1a895 )

Followup: MachineOwner
Hi,
before i forget to ask once again can you please post the make and model of the PSU.

The above dump file means a critical data structure has become corrupted. When the operating system was first installed was it a clean install off a disk/USB or something else?

See if DISM will run after running the sfc, this will attempt to repair any corruption. Open an admin command prompt. Type:
sfc /scannow
press enter and await results

In the same command prompt and after the above scan has finished type:
dism /online /cleanup-image /restorehealth
Press enter and await results (longer this time).

If the first scan found files it could not repair but the second scan is successful, run the first scan again using the same command prompt box and this time it should repair the files found.

Check the event viewer for anything suspicious, sorry i can't be more specific but basically your looking for anything out of the ordinary.
 
The power supply is SeaSonic X-1050 ( SS-1050XM2 )

EDIT: just got a hold of him. He did use a disk for windows 8.1.
 
Last edited:
Its an SSD (Samsung Pro 850 512GB). I assume it was clean. Initially he had windows 10 on but had more severe problems. So he booted from the windows 8.1 disk and presumeably formatted the drive since he had to re-copy his personal files back after installing windows 8.1.

Okay, so I just had him rerun "sfc /scannow" There were problems it couldn't fix.

Immediately following that, I had him run "dism /online /cleanup-image /restorehealth". We received the following message:
The restore operation completed successfully. The component store corruption was repaired.
The operation completed successfully.
Click to expand...

With that being done, we once again ran: "sfc /scannow". This time we got the following message.
Windows Resource Protection found corrupt files and successfully repaired them. Details are included in the CBS.log ...
Click to expand...

So I assume that's what we want. He's gonna be busy for a while today so we'll have to wait till he gets a chance to stress it in some games to see if this had any effect on the crashing.
 
Welp, there was a BSOD today but it was on boot and didn't reoccur upon restart. DMP attached


I am hoping this was a fluke but I somehow doubt it given the history of events!

EDIT:
Additional notes are that late last night games and other 3D programs kept crashing to desktop very quickly. No error logs
 

Attachments

Last edited:
Code: 
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck 50, {ffffc000e28e7d80, 0, fffff8036996836f, 2}


Could not read faulting driver name
Probably caused by : ntkrnlmp.exe ( nt!RtlDeleteNoSplay+5f )

Followup: MachineOwner
Hi,
as this bsod keeps recurring you'll have to basically test everything although after rereading the thread I notice no memtest6 scan was ever run after it was suggested at the beginning'

Windows does have a memory testing app but it can miss errors and the best app for the job is Memtest86.
If you open the link below you'll see you can run Memtest86 in two ways. You can either burn it to disk or install it onto a USB drive it's entirely up to you. You'll then need to enter the bios to change the boot order so you can boot from either the Disk or USB stick you have Memtest86 on.
You must test for at least 12 hours unless it becomes obvious there is a problem straight away.
Memtest86+ - Advanced Memory Diagnostic Tool


Stress test the gpu using furmark:
FurMark: VGA Stress Test, Graphics Card and GPU Stability Test, Burn-in Test, OpenGL Benchmark and GPU Temperature | oZone3D.Net

Monitor system temperatures and run for around an hour unless it bsod's of course.

As a SSD is being used is it fully updating firmware wise and is the controllers set to AHCI?
 
kemical said:
Hi,
as this bsod keeps recurring you'll have to basically test everything although after rereading the thread I notice no memtest6 scan was ever run after it was suggested at the beginning'
Click to expand...

Unless memtest6 is different from memtest86, yes it was done;

Nubkakes said:
Okay, update here.

Memtest 86 ran over the weekend, we got 13 passses with no errors but a warning that the RAM may be vulnerable to high frequency row hammer bit flips.
Click to expand...



We actually used furmark initially when my friend's computer was first built to find out the first GPU was DOA. We couldn't even start the test without it crashing on my friend's computer, so we tested the GPU in 2 known good computers of mine. The first GPU crashed less than 20 seconds in on both good computers. So we RMA'd it and the 2nd GPU(same model/make and same driver version at the time) ran flawlessly for 90ish minutes on both good computers.

I suppose we can try running it for an hour in his computer. If it fails or BSODs maybe we'll get some more hints.

As the for the SSD: we updated the firmware before we assembled his computer because of the rumors and end user reviews warning that certain older firmware versions are known to randomly brick samsung 850 drives. I think he is in AHCI mode but I am not entirely sure. I'll have him check and get back with you.
 
Nubkakes said:
I suppose we can try running it for an hour in his computer. If it fails or BSODs maybe we'll get some more hints.
Click to expand...
It's entirely up to you but it may help.

The other thing you could try is underclocking the gpu. Basically reduce the core and memory speed by a few MHz each time, if the bsod still occurs then drop the speed again. You may find you have to reduce the core and memory by around 100MHz but every card is different so you'll just have to see.
 
Update; he is in AHCI mode. The only other option is RAID
 
Thanks for update. Let me know how the testing goes and good luck.
 
Update: Furmark ran fine for an hour without any changes. It was running just as smoothly at the end as it did in the beginning. No more BSODs since the last DMP I posted. Occasionally some games will run fine for a few hours but then they abruptly start crashing to desktop with no logs. A quick restart of windows remedies this.

We'll see what the rest of the week holds.
 
Ok welI I hope that the coming week goes well and if any bsod's occur please post the dump files and we'll take a look..
 
Well I suppose it was too much to hope for a clear week! 3 more dumps attached.

The first one seems to indicate a DX file, the 2nd one occurred while he was playing World of Warcraft so I figured it might be related, but then Third one occurred when he got up and left his computer alone for a while and I am not seeing anything correlating the 3rd BSOD dmp to the first two.

I am starting to think it maybe the motherboard at this point.
 

Attachments

Code: 
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

Use !analyze -v to get detailed debugging information.

BugCheck 50, {fffff6c80010ee58, 0, fffff8009e3298da, 2}


Could not read faulting driver name
Probably caused by : memory_corruption ( nt!MmDeleteKernelStack+5a )

Followup: MachineOwner
Hi,
I reread the whole thread to try and get a handle on this issue. The only result to come from all the testing is your Memtest86 result:
Memtest 86 ran over the weekend, we got 13 passses with no errors but a warning that the RAM may be vulnerable to high frequency rowhammer bit flips.
Click to expand...

The Passmark web page has this to say regarding the above issue:

PassMark MemTest86 - Memory Diagnostic Tool - Troublingshooting Memory Errors
Do you have any RAM you could lend your friend? Just to see if the issue still occurs after changing the sticks over?
 
Yes, I do happen to have ram modules of the same type/speed/model/brand/size he has. I also know there were no warnings or errors when I let my ram sticks run on my computer with the same motherboard model.

I'll pop a stick out and give in to him to test with.
 
You must log in or register to reply here.

Similar threads

K
  • Solved
Windows 7 BSOD caused by 'ntkrnlmp.exe'
Replies
5
Views
7K
usasma
W
  • Solved
Windows 7 IMAGE_NAME: ntkrnlmp.exe causing BSOD on Win 7 x64
Replies
1
Views
2K
ChatGPT
U
  • Solved
Windows 7 BSOD: ntkrnlmp.exe Please help
Replies
4
Views
3K
kaos
M
  • Solved
Windows 7 BSOD - ntkrnlmp.exe
Replies
7
Views
15K
majklcz
M
C
  • Solved
Windows 7 BSOD ntoskrnl.exe and ntkrnlmp.exe
Replies
12
Views
16K
Yaronmop
Y