[Gigabyte z370n] a study case of BSODs

[alundra]

Dear all,

I'm a little bit out of idea with the below problem, I have spent 10th of hours in it... any help would be really appreciated.

Context and Symptoms:
I am having diverse BSODs (MEMORY_MANAGEMENT,SYSTEM_SERVICE_EXCEPTION,QUOTA_UNDERFLOW...).
The BSOD appears every ~1.5 hours, especially when running games.
The game can be very demanding or not, it the same frequency of errors (i.e : "Path of Exile", "Resident Evil 2 Remake" ...).
Once one crash occurs, the system will BSOD 2-3 times in the next 5 minutes. To work-arond this problem I do often an SFC /SCANNOW, reinstall of nvidia Drivers, and sometimes CHKDSK.

Tests made:
I'm at my 2nd re-install, with CLEAR CMOS (default settings + XMP for RAM), windows 10 Home (UEFI) with the minimum drivers (only windows + Nvidia through GeForceXPerience). Same problem.
I have run memTest86, Prime95 and gsmartcontrol for hours without a crash.
Temperatures are good iddle (35 degrees), and reasonable at load (60-70 degrees).

Configuration:
-MB : Z370N WIFI
-CPU : i5 8600K
-RAM : Corsair Vengeance 2400MHz (in the QVL of the Motherboard)
-SSD M2 : 860 EVO M.2 1TB
-GPU : GeForce GTX 1080 Ti.
-PSU : Seasonic Focus Plus Gold 850W.

Further Ideas:
I have installed BlueScreenView, those are the 4 last MEMORY dumps BSODs (I have more details on the below if needed) :

013119-5531-01.dmp 1/31/2019 9:23:29 PM MEMORY_MANAGEMENT 0x0000001a 00000000`00061941 00007ffa`c6c07ae8 00000000`0000000d ffff958b`56c27a00 ntoskrnl.exe ntoskrnl.exe+1b3440 NT Kernel & System Microsoft® Windows® Operating System Microsoft Corporation 10.0.17763.292 (WinBuild.160101.0800) x64 ntoskrnl.exe+1b3440 C:\Windows\Minidump\013119-5531-01.dmp 6 15 17763 910,596 1/31/2019 9:24:00 PM

013119-5671-01.dmp 1/31/2019 8:39:58 PM SYSTEM_SERVICE_EXCEPTION 0x0000003b 00000000`c0000005 fffff802`4411445b ffff8483`d38ee770 00000000`00000000 FLTMGR.SYS FLTMGR.SYS+44be Microsoft Filesystem Filter Manager Microsoft® Windows® Operating System Microsoft Corporation 10.0.17763.1 (WinBuild.160101.0800) x64 ntoskrnl.exe+1b3440 C:\Windows\Minidump\013119-5671-01.dmp 6 15 17763 1,114,732 1/31/2019 8:40:33 PM

013119-5656-01.dmp 1/31/2019 8:32:41 PM QUOTA_UNDERFLOW 0x00000021 ffffaa8f`c2edd080 00000000`00000002 00000000`fffffef4 00000000`fffe6c79 ntoskrnl.exe ntoskrnl.exe+1b3440 NT Kernel & System Microsoft® Windows® Operating System Microsoft Corporation 10.0.17763.292 (WinBuild.160101.0800) x64 ntoskrnl.exe+1b3440 C:\Windows\Minidump\013119-5656-01.dmp 6 15 17763 562,292 1/31/2019 8:33:56 PM

013119-5625-01.dmp 1/31/2019 7:35:10 PM SYSTEM_SERVICE_EXCEPTION 0x0000003b 00000000`c0000005 fffff804`59fb83ea fffff88e`be334e30 00000000`00000000 dxgkrnl.sys dxgkrnl.sys+1183ea DirectX Graphics Kernel Microsoft® Windows® Operating System Microsoft Corporation 10.0.17763.194 (WinBuild.160101.0800) x64 ntoskrnl.exe+1b3440 C:\Windows\Minidump\013119-5625-01.dmp 6 15 17763 1,724,380 1/31/2019 7:35:44 PM


Questions:
1. could the problem be related to a compatibility issue of software/drivers OR hardware ?
2. could the problem be the BIOS (I've flashed the BIOS to F10 version) ?
3. could this be related to the PSU ?

 

[Traditore]

Google for a previous GPU driver version and install it without GF Experience.

Download https://sdi-tool.org/ and update all the drivers except the GPU driver.

If you can borrow a PSU or a GPU from a friend - try swapping them one at a time, that way you find the culprit by process of elimination.

 

[alundra]

Google for a previous GPU driver version and install it without GF Experience.

Download https://sdi-tool.org/ and update all the drivers except the GPU driver.

If you can borrow a PSU or a GPU from a friend - try swapping them one at a time, that way you find the culprit by process of elimination.

thanks for proposing.

1. unfortunately, I cannot install manually an nvidia Driver without GeForceXPerience, it's saying it's not compatible with this version of windows (which is a Windows Home with a proper Key). I googled it, it's kinda of normal for windows UEFI installation.

2. I will try this.

3. that would be the best shot I guess, not everyone has a spare GPU, I'll try this.

4. I just noticed that gigabyte published (10 days ago) a new BIOS version "F11" saying :
Update CPU microcode
Improved system performance and stability
-> I will also try this.

---------------------------------------------------------------------------------------
[03.02#13:00]
I had bluescreens, so many and so quick that I can't use windows anymore. I will have to re-install anyway.

Now I'm playing with UBCD, I have clear CMOS (again :/ ) this time I didn't put the XMP profile for the RAM.

---------------------------------------------------------------------------------------
[03.02#19:32]
I have run for 1:20 hour memTest on each RAM stick individually, on each slot individually.
- one is giving errors on tests #7 and #9 and "with a confidence value 253 (which is very high).
- one is having no errors at all.

-> as I have already replaced the RAM with garantie due to RAM errors, it is such a coincidence the same RAM (brand and model) is also having an issue, could it be the MB or PSU are killing the RAM stick !? could it be due to high temperatures on the MB ? (the sensors do not allow to know the temperature everywhere, but with HWmonitor I noticed reasonable ones.)

---------------------------------------------------------------------------------------
[07.02]
The problem was the RAM, I replaced it and now it's working like a charm.
Thanks all for the support. Good day.

 

[alundra]

Google for a previous GPU driver version and install it without GF Experience.

Download https://sdi-tool.org/ and update all the drivers except the GPU driver.

If you can borrow a PSU or a GPU from a friend - try swapping them one at a time, that way you find the culprit by process of elimination.

new information.

I had bluescreens, so many and so quick that I can't use windows anymore.

I have run again a couple of memTests, some found nothing, and once it found 1000s of errors during the Test #7 block moves, so much errors that the memTest stopped.

Last week I have changed the RAM (same model and brand) with the garantie, as I was detecting errors with Windows Memory Diagnostic Tool.

But looks like something wrong here again, could it be the PSU kind of corrupting the RAM or the MB slots ?

I'm running the default BIOS RAM configuration (Auto), just with the profile XMP "default 1", which is in the QVL testing PDF...

I have UBCD, so I will continue my tests, but if you see some to prioritize please let me know.

thanks.

 

[kerberos_20]

disable XMP, run test again

 

[alundra]

disable XMP, run test again

[03.02#13:00]
I have removed the XMP default 1 profile, load optimized configuration from the BIOS (hence RAM should be default configuration).

I've run the memtest86 v4.3.7 from UBCD, 2 times.
- once nothing.
- the second, got :
-- 2227 errors in Tests #7.
-- 65K+ errors in test #8.

[03.02#16:20]
-> I've launched the test on one RAM stick and got errors on test #7 on both slots. (getting the error after 20 minutes of tests)
-> I'm testing a.t.m with the other RAM stick.