BSOD Thread_Stuck_In_Device_Driver when playing games with Ryzen 4800U

Maetel · 16 Jul 2020

Hello!

I'm posting here as I have been battling with my new laptop for a month now without success. My current configuration is:

Lenovo IdeaPad 5-15ARE05

CPU: Ryzen 7 4800U
GPU: Integrated Vega Graphics
RAM: 16GB DDR4 3200Mhz Hynix dual channel
SSD: 512GB NVMe TLC Samsung MZVLB512HBJK
Windows 10 Home 64-bit Version 2004 build 19041.388

I have received this laptop on 04th of June and it works as it should for almost everything I need it for. The only time it crashes with a THREAD_STUCK_IN_DEVICE_DRIVER BSOD is when I play games. It started from the very first day I got the laptop as I installed Steam straight away, and it occurs on every game I have tried so far. Such games are:

A Bird Story, Castle of Illusion, Game of Thrones, Tales from the Borderlands, The Walking Dead: Michonne, The Wolf Among Us, Valiant Hearts

If any of you is a gamer and knows these games you should know they are not very demanding and my machine is perfectly able to run them at max settings. Such crashes are "random" and there is no specific point in a game where the system will always crash.

What I have done so far in order to find what's causing instability is the following:

- using the laptop monitor or an external one, but no matter which one I use the BSOD will appear
- reinstalling W10 with Version 1909, because the 2004 one is known to have issues with Lenovo laptops. Unfortunately yesterday the system has crashed anyway...
- disabling Turbo Boost as this laptop gets very hot when under load. So hot that it would reach 100C everytime, leading me to believe the system was simply shutting down due overheating. With the CPU running at the stock 1.8GHz and not reaching 4.3GHz anymore the maximum temperature reported by both HWInfo and Hardware Monitor is 60C which is very reasonable, but yesterday I still got the BSOD, so supposedly it's not a temp problem either...
- trying two different sets of GPU drivers. The first one is provided directly by Lenovo here: laptops and netbooks :: 5 series :: ideapad 5 15are05 - GB and the second one by AMD here: https://www.amd.com/en/support/previ...-ryzen-7-4800u and when I did the latest one was Adrenalin 20.5.1.
I have not tried the latest version released last week as I have had issues with already 2 different drivers, and "apparently" the ones provided by the OEM are more stable with this particular machine
- increasing VRAM from 512MB to 2GB in the BIOS, even though I know W10 should allocate it dynamically
- playing both in full screen or windowed, no change here too
- running MemTest86 overnight with 8 passes but no error message
- running full diagnostics with the Lenovo Diagnostics Windows app, however I will probably try things like CrystalDiskInfo, etc. to test again the components
- using sfc /scannow, dism /online /cleanup-image /scanhealth, dism /online /cleanup-image /restorehealth and chkdsk /scan as I saw them mentioned here on another thread, all good tho

I thought that this was due overheating / buggy W10 / buggy GPU driver but it seems they are not the cause, so now I'm trying different things unless you can point me to the right direction. I have today updated the BIOS as a new version has been released and changed TDR Delay by adding a voice in the registry and using the value "8" as suggested on various websites. If nothing works I will see if these issues happen whether the laptop is running on battery and/or on mains, but after that I will not know what else to do.

I'm obviously uploading the V2 thingy you ask for and attach a link to Google Drive for one of the memory dumps.

Thanks and let me know if you need anything else!!

MEMORY.DMP - Google Drive

philc43 · 17 Jul 2020

Do a factory reset and if it still has BSODs then return under warranty. Overheating can often trigger a BSOD.

The memory dump indicates the problem was related to the graphics system and drivers. On a new system I would always use the Laptop Manufacturer provided drivers as these will be optimised for the system.

Maetel · 17 Jul 2020

Hi, thanks for looking at the memory dump for me. To be honest yesterday I had forgotten to say I had analysed it with WinDbg preview and noticed that amdkmdag.sys or dxgkrnl.sys where involved, but I thought that file can give you more information.
I have done already a factory reset and as I said it's not an overheating issue as this happens even when the CPU / GPU run at 60C while gaming. I don't know if Youtube / Disney+ / Media Player Classic / etc. use DirectX to show videos, but when I use them nothing happens.
Unfortunately I'm not in the position to return the laptop as I work from home due to the pandemic. My 8 years old desktop has died the very same day I received this machine, so I'm waiting for AMD and NVidia to release new CPU / GPU before assembling a new desktop, which will take probably 3-4 months...
Also it took over a month for Lenovo to dispatch and they clearly told me again that if I were to return it to them it would take 1-2 months before getting it repaired / replaced.
So, what other kind of troubleshooting can I do? This happens only when I try to play games, so I may have to give up on gaming for now, but I would like to at least know what's causing this.

Thanks

philc43 · 17 Jul 2020

Try running the AMD cleanup utility and then reboot and let Windows install drivers as needed - they should all automatically be picked from the Windows Update database.

Maetel · 17 Jul 2020

Hi again, yes I will do that tonight. I had already used DDU two days ago but I heard it may not work properly sometimes and the utility you mention is better.
Also since you're here... when I reinstalled W10 Windows Update installed Advanced Micro Devices, Inc. - Display - 26.20.14003.2003, and that's the version I can see when I look at Device Manager. However the driver I have installed from Lenovo is 26.20.14001.580 and I don't understand if there are 2 drivers installed at the same time or not, I cannot even uninstall the first one from Windows Update...

Thanks

Ztruker · 18 Jul 2020

A new laptop should be having the problems you are encountering. You should definitely start the RMA process and get on them about the 1-2 month crap. There should be no more than a 1 week turn around time.

Tell them you arer not interested in them repairing it at this time, just replace it .