Drivers & Software

[maetel]

Greetings,

I have recently purchased a Lenovo IdeaPad 5-15ARE05 laptop which works flawlessly when I work but constantly crashes when I play games on Steam. My current configuration is:

Ryzen 4800U with Vega graphics
16GB DDR4 3200MHz SK Hynix
512GB SSD NVMe Samsung MZVL B512HBJK
Windows 10 Home 64-bit ver.2004 up-to-date
AMD VGA_26.20.14001.58006_WHQL driver (also tried Adrenalin 2020 Edition 20.5.1)

In short only when I play games on Steam the laptop will randomly freeze, regardless of the game, and a BSOD will appear due THREAD_STUCK_IN_DEVICE_DRIVER caused by amdkmdag.sys or dxgkrnl.sys.
I'm not sure what the problem exactly is but it seems to be software based related to the GPU which makes sense as this happens only when I play.
This crash happens whether I use an external monitor or the internal one and it has appeared with both the latest W10 version and the previos v.1909 one, so it's unlikely to be caused by Windows.
I have also disabled the Turbo boost as I noticed the laptop was overheating and I would see 100C in both HWInfo and Hardware Monitor Pro, but even now that the CPU barely reaches 60C when playing without performance drops the laptop still crashes, so it's not caused by the high temperatures either.
As I have also tried two different drivers, one taken from the Lenovo website and one from AMD, it seems they may not be the culprit.
I have ran Memtest86 overnight with 8 passes but no error message and ran diagnostic provided by Lenovo, but again no issues were reported.
I'm running out of options and I was wondering if you could help me in finding out where the problems and how to fix it.

Cheers

[albertfu]

I also have the same BSOD THREAD_STUCK_IN_DEVICE_DRIVER issue with my new Ryzen 4800H laptop (without discrete GPU, so it is more or less the same configuration as the 4800U one in your post).

By analyzing .dmp files, it also points to the same files that you mentioned (0xEA_IMAGE_dxgkrnl.sys，atikmdag.sys). 

BSOD happens randomly when any of hardware monitoring tools are running (HWiNFO, AIDA64, GPU-Z), even without any 3D or video encoding/decoding workload.

It also happens sporadically when videos are being played with MadVR (so integrated Vega GPU is heavily used) or during some synthetic 3D workloads like FurMark or Superposition benchmark.

I didn't try playing games on this laptop, but I am sure it will reproduce the BSOD issue, since it tends to occur whenever the Vega iGPU is being loaded.

I tried to run Memtest and other stability tests on this laptop but could not find issue other than the BSOD 0xEA related to the Vega iGPU.

The APU's temperature is absolutely fine, never exceeds 85C even under heaviest load, and the BSOD can occur even when the APU is almost idle (just opening GPU-Z while doing nothing can sometimes cause BSOD), so I am sure it's not temperature related.

Same as yours, BSOD occurs regardless internal or external display is being used.

I tried to install 20.5.1 version Radeon driver from AMD site (uninstall the old one, then reboot into safe mode and run DDU, reboot, install the new one), it is more stable and newer then the OEM one but BSOD still occurs.

I will try the latest version 20.7.2 when I got time, since I highly doubt it is related to immature driver/BIOS.

Last thing to mention, I already did 2 RMA with this laptop, and BSOD 0xEA occurs on all 3 of them (all of them arrived brand new and sealed).

Hoping someone could bring a fix to this issue or at least find the real culprit, I really love the perf/watt or perf/weight of Renoir laptops and this issue is preventing me from ditching my Intel laptop.

Best regards

[maetel]

Hi albertfu, thanks for your reply. Do you have an IdeaPad too?

To be honest your case seems to be worse because I'm keeping HWInfo and Hardware Monitor open at all times but there are no BSODs unless I play.

Since my original post I haven't had the chance to test much and I noticed that actually I have always been with some drivers provided with Windows Update, and if I try to install the ones provided by Lenovo nothing will happen apart from installing Radeon Lite, so I will check those later...

Also I have focused more on testing my hardware now, but:

-SeaTools with long generic test doesn't show any errors on the SSD

-Prime95 shows no errors on any threads after 3 hours of testing wiith my CPU running at a stable 70-75C

-Furmark (unlike for you) runs flawlessly even after 2 hours with all the options enabled and the GPU working between 60-70C

So now I'm even more puzzled as I keep being told to simply ask for a replacement, but I'm pretty sure the same thing that happened to you would happen to me.

Now I have updated the BIOS as a new version has been released last week, disabled Windows Game Mode as it may give problems while gaming to some people and added an option in the registry to increase the TDR delay to 8-10s, even if I don't think it will help...

After that I will make sure the drivers from WU are completely erased and the ones from Lenovo are installed, after that I will install the ones from AMD. ANd if still the laptop crashes I will check if this happens when it runs on battery too.

After that... no idea lol, tbh I'm "happy" (not for you) you replied as I feel more people are affected but noone says anything.

[albertfu]

Hi maetel, I got a Mechrevo laptop with Ryzen 4800H, so a different brand and model, but the basic configuration is the same: Renoir APU without discrete GPU.

Regarding the graphics driver, my personal experience is that the 20.5.1 version is much more stable than the one shipped with the laptop: 19.40.52 (26.20.14052.0).

Actually, I had very few BSOD 0xEA ever after installing 20.5.1, most of them were when I checked sensor page while doing almost nothing. FurMark or Superposition have not crashed ever since that. So, it is really sporadic and annoying.

I will test the latest version 20.7.2 later and see if things get better or worse.

On the hardware side I didn't push very hard on the stability test, like running P95 small FFT.

I just used AIDA64 (tried various combinations of FPA, CPU, cache, memory stress tests), looped Cinebench R20, and Asus RealBench for stressing the APU.

Never ever it had any issue when GPU isn't involved, and the temperature was always fine (APU power consumption around 45 watts and temperature under 85C).

Also, the 0xEA BSOD occurs on all 3 of brand-new laptops when GPU driver is not updated, so I guess the issue is really on the driver/BIOS side. The chance that I got 3 faulty laptops in a row is really low.

I found very few reports of 0xEA BSOD on Ryzen 4800U/4600U machines on the web, most of them (tenforums, computerbase.de) are about Lenovo IdeaPad, also some not very detailed report about Lenovo Xiaoxin 15 (Chinese equivalent of Ideapad I assume).

But rest assured it's not an isolated case and what we need to do is to get some attention from AMD so they can start investigating this. I can provide some minidump files if they need them.

Best regards

[albertfu]

Hi maetel, just to confirm that the same 0xEA BSOD still occurs with version 20.7.2, but it's quite rare.

I didn't figure out a scenario where it can be systematically reproduced yet, but it didn't occur when GPU-Z or HWiNFO are not running in the background. 

[maetel]

Hi, sorry to hear that. I think I have instead fixed my laptop as I have accumulated around 45 hours of gaming now without a single crash, so for me this issue is solved.

I don't know what exactly has fixed it, but as I have said last time I have:

-disabled Windows Game Mode

-updated BIOS to LENOVO E7CN26WW 22/06/2020 (don't think it would be the same one for your laptop)

-increased TDRdelay to 8s by adding a voice to HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Control\GraphicsDrivers  Value Name: TdrDelay and Value Data: 8

-used AMD Cleanup Utility to erase every driver and app related and let W10 install the drivers Advanced Micro Devices, Inc. - Display - 26.20.14003.2003 (again, not sure it they would be the same for your machine)

Given you seem to experience the BSOD in different circumstances I'm afraid I can't help more than this, but good luck!