[SOLVED] Windows 10 BugCheck error - Page fault in nonpaged area

[Colif]

Last error is actually a more common AMD driver error, I see Thread stuck in device error and I know its a AMD GPU. Which GPU was in PC for last BSOD?

TDR is an abbreviation for Timeout, Detection, and Recovery. It's a windows component. Whenever you receive this error, the Graphics card stops responding and Windows will stop to fix the driver error.

Its obviously not the card. I am wondering what else might be causing it.

Asus Prime Z490-P MBO
Intel Core i5-10600 CPU - On Asus QVL list
Corsair VENGEANCE LPX 16 GB kit(2x8GB) DDR4 DRAM 3600MHz PC4-28800 CL18, 1.2V / 1.35V, CMK16GX4M2D3600C18 - NOT on Asus QVL list
Samsung 970EVO NVMe M.2 SSD 500GB - System Drive
SSD Crucial P1 500GB 3D NAND NVMe PCIe M.2 SSD
Corsair 650W PSU - RM650 Fully modular
MSI Armor RX570 graphic card

anything old attached to PC? USB drives for instance, USB devices are linked to PCIE via controllers so its always possible a old mouse or something is causing the GPU drivers to stop working. Long shot, I had a mouse causing my GPU drivers to stop working a few years back and I always think of it if the obvious answer isn't the answer,

 

[nikicam]

Last error is actually a more common AMD driver error, I see Thread stuck in device error and I know its a AMD GPU. Which GPU was in PC for last BSOD?

TDR is an abbreviation for Timeout, Detection, and Recovery. It's a windows component. Whenever you receive this error, the Graphics card stops responding and Windows will stop to fix the driver error.

Its obviously not the card. I am wondering what else might be causing it.

Asus Prime Z490-P MBO
Intel Core i5-10600 CPU - On Asus QVL list
Corsair VENGEANCE LPX 16 GB kit(2x8GB) DDR4 DRAM 3600MHz PC4-28800 CL18, 1.2V / 1.35V, CMK16GX4M2D3600C18 - NOT on Asus QVL list
Samsung 970EVO NVMe M.2 SSD 500GB - System Drive
SSD Crucial P1 500GB 3D NAND NVMe PCIe M.2 SSD
Corsair 650W PSU - RM650 Fully modular
MSI Armor RX570 graphic card

anything old attached to PC? USB drives for instance, USB devices are linked to PCIE via controllers so its always possible a old mouse or something is causing the GPU drivers to stop working. Long shot, I had a mouse causing my GPU drivers to stop working a few years back and I always think of it if the obvious answer isn't the answer,

Hi Colif,
thanks for your continuous support on this topic!

During that last error on my new PC, I had an RX550 graphic card installed in my system.

1. I have put my RX550 card from my old PC to the new one i5 10th-gen and run the OCCT stress test. In two tests the system froze both times. The first time the system became unresponsive so I had to reset the machine using the reset button. After PC booted again I found this in my system log:

Event 1, WHEA-Logger (A fatal hardware error has occurred. A record describing the condition is contained in the data section of this event. During this freeze, the dump was not saved.

The second freeze came up after only 1 minute of the OCCT stress test with the following dump error posted in the previous post with TDR error.

I am using Logitech MX Master 3 mouse attached to USB, nothing else is attached to the USB controller. I am also using the Logitech Bluetooth keyboard, but I doubt that the keyboard could be the culprit.

My friend still suspects the memory or even a fault Motherboard. That sounds unlikely to me as my system passed memtest with XMP profile ON two times and 8 hours in total without any errors.
Also without the GPU and AMD driver, my new system passes OCT tests without any issues.

 

[nikicam]

Sorry, I've noticed that I had 3 dump entries that happened when I was using RX 550 graphic card in my system.
As you can see below, the new one includes dxgkrnl.sys module also.

On Sat 16-Jan-21 10:36:52 your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\011621-5546-01.dmp
This was probably caused by the following module: dxgkrnl.sys (dxgkrnl+0x4289D)
Bugcheck code: 0x100000EA (0xFFFF858E0A2FA0C0, 0x0, 0x0, 0x0)
Error: THREAD_STUCK_IN_DEVICE_DRIVER_M
file path: C:\Windows\system32\drivers\dxgkrnl.sys
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: DirectX Graphics Kernel
Bug check description: This indicates that a thread in a device driver is endlessly spinning.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem.
The crash took place in a Microsoft module. Your system configuration may be incorrect. Possibly this problem is caused by another driver on your system that cannot be identified at this time.

On Sat 16-Jan-21 10:36:52 your computer crashed or a problem was reported
crash dump file: C:\Windows\MEMORY.DMP
This was probably caused by the following module: amdkmdag.sys (amdkmdag+0x70C00)
Bugcheck code: 0xEA (0xFFFF858E0A2FA0C0, 0x0, 0x0, 0x0)
Error: THREAD_STUCK_IN_DEVICE_DRIVER
file path: C:\Windows\System32\DriverStore\FileRepository\u0361901.inf_amd64_204a65b18f2a904a\B361909\amdkmdag.sys
product: ATI Radeon Family
company: Advanced Micro Devices, Inc.
description: ATI Radeon Kernel Mode Driver
Bug check description: This indicates that a thread in a device driver is endlessly spinning.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem.
A third party driver was identified as the probable root cause of this system error. It is suggested you look for an update for the following driver: amdkmdag.sys (ATI Radeon Kernel Mode Driver, Advanced Micro Devices, Inc.).
Google query: amdkmdag.sys Advanced Micro Devices, Inc. THREAD_STUCK_IN_DEVICE_DRIVER

On Sat 16-Jan-21 10:36:27 your computer crashed or a problem was reported
crash dump file: C:\Windows\LiveKernelReports\PoW32kWatchdog-20210116-1036.dmp
This was probably caused by the following module: ntkrnlmp.exe (nt!IoUnregisterShutdownNotification+0xF84)
Bugcheck code: 0x1A1 (0xFFFF858E0A32C040, 0x0, 0x0, 0x0)
Error: CUSTOM_ERROR

The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.

 

[Colif]

dxgkrnl.sys = Direct X. Its been in all of them, If I see it crash almost every time its caused by gpu drivers. Its not telling us anything new.

So 550 produces normal AMD GPU errors.
the 570 produces abnormal ones but they point at drivers.
cards work fine outside of PC...

So your intel system hates AMD.

ram isn't on QVL but ram errors don't stick to 1 driver, they bounce around. Often if you replace a driver, the BSOD will move to another, not stay on same driver again.
I wonder if it will blame Nvidia drivers if you put an Nvidia card in.

Its something else , just what. I am not exactly sure. One of your other parts of the PC keeps pushing GPU drivers to crash.
Could be PSU
Could be motherboard
I doubt its the SSD/storage

they all work fine without a GPU in PC... which makes it hard.

have I suggested a repair shop? really curious what it does with a Nvidia card.

 

[nikicam]

I have reported this to AMD and will wait to see what they'll have to say about this.
Unfortunately, I don't have an Nvidia card to test, but yeah, that was my plan. Hopefully, my system just doesn't get along with the AMD drivers and products.

Well, OCCT stress tests the PSU too, and considering that RX550 doesn't require external power, but it powers itself from the PCI-E slot and I still get the freeze.

It's not the SSD NVME storage since I have changed tried both with Samsung 970EVO NVMe M.2 SSD and SSD Crucial P1 500GB 3D NAND NVMe. I got freezes in both cases.

Could it be just that my new system needs an updated driver from AMD that could fix this issue or I have a faulty hardware component no matter that the tests are passing fine without GPU installed?

 

[Colif]

Could it be just that my new system needs an updated driver from AMD that could fix this issue or I have a faulty hardware component no matter that the tests are passing fine without GPU installed?

Something isn't right, I don't think its a driver, not after 2 gpu and multiple driver versions. You run ddu each time you reinstall drivers?
I am at the throw anything at wall and see if it sticks stage. Maybe I ask if anyone else has any smart ideas

I assume you clean installed win 10 at some stage?
new beta bios for motherboard released 15th Jan.

PRIME Z490-P BIOS 1602
"Improve system performance and stability
Improve ReSizable BAR compatibility for NVIDIA RTX30 series graphics cards

small chance bios fixes problem, its such an odd problem.

 

[nikicam]

I have used clean system install twice using both Win 10 2004 and Win 10 20H2 version.
One thing I have noticed is that once I did a clean install of Win 10 2004 I haven't had issues for approximately 2 months after which I have upgraded to 20H2 and the freeze issues cam back.
Maybe I just had luck during that time frame and went without a restart.

I've just noticed that bios v1602 isn't available anymore. Did they just removed it??

PRIME Z490-P｜Motherboards｜ASUS Global

 

[Colif]

i have asked if anyone else can think of what I might have missed.

 

[nikicam]

I've just used OCCT to stress test the PSU, but since it's also using GPU, at least I think it is, it crashed again with the same error as before:
On Sat 16-Jan-21 16:01:23 your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\011621-7406-01.dmp
This was probably caused by the following module: amdkmdag.sys (0xFFFFF8053C23BA43)
Bugcheck code: 0x50 (0xFFFFA98050346F12, 0x2, 0xFFFFF8053C23BA43, 0x2)
Error: PAGE_FAULT_IN_NONPAGED_AREA
file path: C:\Windows\System32\DriverStore\FileRepository\u0361901.inf_amd64_204a65b18f2a904a\B361909\amdkmdag.sys
product: ATI Radeon Family
company: Advanced Micro Devices, Inc.
description: ATI Radeon Kernel Mode Driver
Bug check description: This indicates that invalid system memory has been referenced.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem.
A third party driver was identified as the probable root cause of this system error. It is suggested you look for an update for the following driver: amdkmdag.sys (ATI Radeon Kernel Mode Driver, Advanced Micro Devices, Inc.).
Google query: amdkmdag.sys Advanced Micro Devices, Inc. PAGE_FAULT_IN_NONPAGED_AREA

Here are the screenshots of OCCT setup and test:
https://ibb.co/dgzT6Pj

power-stress-test

Image power-stress-test hosted in ImgBB

ibb.co

 

[nikicam]

Hey Colif!!!
we moved a step further. I have disabled the XMP profile in BIOS and the system passed the OCCT test without freeze!!

I can see now that bios 1602 is available again. I'll try with that and with XMP ON and run the OCCT test again.

Update: After BIOS update to 1602 and XMP ON, the system crashed again in the OCCT test.

Ok, who's to blame now, Corsair memory or Asus motherboard? Both hardware is probably fine but just don't work well together.

 

[Colif]

Corsair VENGEANCE LPX 16 GB kit(2x8GB) DDR4 DRAM 3600MHz PC4-28800 CL18, 1.2V / 1.35V, CMK16GX4M2D3600C18 - NOT on Asus QVL list

maybe neither since its not on list. Asus would point at their listing... Corsair would point at fact it worked for them.
Isn't this how we started this thread? You in the middle. I mostly ignored it back then as I used to think ram needed to be on QLV to work but I was told that was not the case. So I figured, Corsair got it to work so it should. Now I see there is more grey area.

You might be best grabbing ram tested by Asus and then you would have a starting point.

One suggestion I had about current situation is - Do a clean install of Windows as your PC dislike for AMD cards could be a registry problem.

Right now, ram is more likely than windows since everything has to run in it. If its just a weird incompatibility between your sticks timings and the MB, that might be enough to have it always blame same driver. It goes against everything i normally see with ram errors but I can't be right all the time...

 

[nikicam]

Yeah, if that would be simple.
For instance, this memory is on Asus QVL:
CMK16GX4M2B3200C16

But, Asus only supports version 4.24 with Samsung B-die and there's also a version v5.39 with SK Hynix. Since you cannot choose the memory type by the version number but only the part number I'll probably get v5.39 with Hynix B-die.
I have checked with Asus regarding the version number and they have told me that they ONLY support v4.24 that is on the QVL, but not v5.39 even though this memory has the same part number. So, it will be very hard to buy the memory that actually is on Asus QVL.

I've found only this one CMK16GX4M2A2666C16 that is on Asus QVL and costs 100€, but I'm not sure that I should buy that since my memory runs at 2133 without XMP and this one runs at 2666 with XMP but has lower latency 16 vs 18 (not sure that I would feel that). I really don't think it's worth it to sell my current memory and buy the slower one just to get that 500Mhz jump.

Here's one interesting comment regarding CMK16GX4M2B3200C16 on Newegg:

Cons: Expensive Apparently there are two different versions of CMK16GX4M2B3200C16. ver4.xx with Samsung B-die AND ver5.39 with SK Hynix. For the price, I expected Samsung b-die but instead got ver5.39 with SK Hynix - VERY DISAPPOINTED Product description needs to state which version you are paying for. I never would have ordered this had I known what I was getting. Stable one minute, not so much the next.

Here's my conclusion:
no matter that this memory is not on the QVL from ASUS it definitely should work at rated speeds without issues. Check this out:
CMK16GX4M2B3600C18 - on ASUS QVL
CMK16GX4M2D3600C18 - not on ASUS QVL

The first one comes with a heat spreader and I presume that's the only difference.

Also, here's a reply I got from a Corsair employee regarding my issue on Corsair forum:

Hey NiK,

I have a suggestion! I would do this if I had the exact same issue:

1. Reflash latest bios ----Go to Step 2 if this still gives issues.

2. Reflash previous bios version ----Go to Step 3 if this still gives issues.

3. Start an exchange.

While your kit may not show up on the Asus QVL, I wouldn't think the kit would struggle to run at 3600MHz on Z490 without issues.

Anything else, feel free to reach out our support team: support.corsair.com

-Art

According to this post, my memory uses B-Die which should be good. I have also checked the label on the heat spreader and I can see version 4.32.
CMK16GX4M2D3600C18 not doing 3600MHz on my board | Overclock.net

 

[Colif]

But, Asus only supports version 4.24 with Samsung B-die and there's also a version v5.39 with SK Hynix. Since you cannot choose the memory type by the version number but only the part number I'll probably get v5.39 with Hynix B-die.
I have checked with Asus regarding the version number and they have told me that they ONLY support v4.24 that is on the QVL, but not v5.39 even though this memory has the same part number. So, it will be very hard to buy the memory that actually is on Asus QVL.

You can't blame Asus if its Corsair that uses same part number and has 2 different types of ram. I didn't know I had SK Hynex myself until I saw CPU Z recently. I didn't go looking for B Dies as well, I got PC in 2020, its not ideal time to be picky. That and low latency ram was also stupidly expensive, at same time I got 32gb of cl16, I could get 16gb of CL14 for twice the price.

CMK16GX4M2B3600C18 - on ASUS QVL
CMK16GX4M2D3600C18 - not on ASUS QVL

There is new ram made all the time, the makers of motherboards don't test all of it and add them to new motherboards, but as the only difference between those 2 sticks is a B becomes a D, its possible that the later is just an updated version of the former. As time passes it becomes harder to find ram that is on list as most is replaced by ram maker.

Ram doesn't run in XMP without causing crashes. Is the speed of the ram that much slower to warrant replacing it as if PC actually works without BSOD otherwise, I would take it as is.

 

[nikicam]

Ram doesn't run in XMP without causing crashes. Is the speed of the ram that much slower to warrant replacing it as if PC actually works without BSOD otherwise, I would take it as is.

You are absolutely right Colif and I agree with you, but it hurts when you pay for something that you can't use, although I seriously doubt I'll notice the difference. But then again, I could just spent less and buy CMK16GX4M2A2666C16 and run it with XMP without issues.
Also, one question, would I be able to tell the difference between 3200 C16 and 3600 C18?

I have opened a ticket with Corsair and asking them for advice. I don't plan to RMA just yet.

Colif, thanks a lot for your kind help! I will update this topic once I receive a reply from Corsair.

 

[Colif]

I agree about the paying for something you can't use. I would be just as annoyed as you could be. I am not sure what I would do. But since you spent last 2 months fighting a GPU that wasn't the problem at all, I would take the loss as now you can use PC as you wanted to.

Also, one question, would I be able to tell the difference between 3200 C16 and 3600 C18?

maybe on benchmarks, i don't think it is as lot in reality, maybe a few frames in games.

 

[nikicam]

Corsair wants me to clear motherboard CMOS to factory defaults and test each stick individually with Memtest86 with BIOS. Not sure why they want this as I've already said that I already did two complete tests with Memtest86 with both sticks and XMP profile ON without errors. The problem is that the memory won't work in Windows with XMP profile at the advertised speed at 3600Mhz.

I've never heard that someone had successful Memtest86 passed in dual mode and then had errors when sticks were tested individually.

 

[Colif]

i thought you were just staying with what you had?

why test ram if you keeping it?

 

[nikicam]

i thought you were just staying with what you had?

why test ram if you keeping it?

As said before, I asked Corsair for advice and explanation on why this memory won't work at the advertised speed.

 

[Colif]

oh okay.

up to you if you test ram individually. I don't know if outcome can be different if tested alone. I would be surprised if it is. Normally want to test 1 stick at a time just so you can identify the bad ram. I wouldn't expect a different outcome from when you tested both.

 

[Colif]

It was suggested you could run xmp with a slightly higher dram voltage, see if that helps.