[SOLVED] Please help me, I get "TDR error, atikmdag.sys" +black screens after 10 mins of gaming on a New build.

Feb 8, 2020
9
0
10
Watch the language
Type: Desktop
GPU: No

CPU: Ryzen 3 2200g with integrated Vega Graphics.

Motherboard: MSI A320m pro vd/s v2

RAM: Corsair Vengeance LPX 2400mhz 4gb ×2

Operating System Version: Windows 10 Pro 1903.

Hello, I get that notification in the notification center and I get a black/blue screen every time I click on the start button or anything windows related, but the processes still run (say I was watching a film, and I can still hear the music even after black screen).

I get the display back after I press win+ctrl+shift+B, what it apparently does is reset the graphic drivers, but the black or blue screen occurs again after a second and is super annoying. I tried installing all the driver versions (Vega 8), the problem remains the same.

When I use an Application like a game that is barely resource intensive, it gets stuck and says, "TDR error, atikmdag.sys" and restarts.

And, when it does that and I login to win10 and check the notif center I see that both the Application and the Windows operating system itself was apparently blocked access to Graphics hardware.

I must mention that my PC is new, I have memtested the rams, scanned the system, reinstalled win 10 thrice (once with internet). I am using the processor graphics itself so there must be no problem with the hardware.

I don't get that error in safemode or if I uninstall the drivers, but please don't tell me that it's probably because of wrong drivers coz I had a ticket regarding this issue with AMD and they sent me a couple of links to the drivers.

No overclocking. Tried disabling global cstate option, increasing soc voltage, increased the tdr delay to 8 but that causes video scheduler internal error.

The random black or white screen (depending on the windows app's color I was using) which goes away after pressing win+shift+b+ctrl.

Black screen with all the processes still running

White screen

Also, I should say that:-

I tried my tv as a monitor and that didn't change things.

The radeon v20 softwares especially caused a reboot loop while doing anything without showing the bsod unlike v19.

Ran furmark test for 1 hour the max temp. was 57 °C no bsod.


Please help me.

SOLVED!!

I solved the problem by getting a replacement of the ram, because it was faulty. Not sure why Memtest86 said 0 errors tho. I got the exact same model of ram, now everything is fine. It seems like the other set of ram was sending corrupted data, hence the black screens.
 
Last edited:
Solution
Have you enabled XMP settings in UEFI, or whatever MSI refers to these parameters for Ryzen. I'd also recommend Memtest86 to check RAM, as the iGPU uses system memory to function on these APUs. A small bump to memory voltage could help. I had to do exactly this to fix crashes/display corruption on a build I setup for family with a Ryzen 2200G.
Okay, thats quite especific, sounds to me like when I overclock my GPU, i get driver error/crash, full black screen with background going indicating that only the video has crashed, wich sorta looks like your case.

well, by what ive heard there, 1 could have been driver issues, but you installed, reinstalled no OC, everything is fresh and all, and still no avail? im gonna risk saying a probable integrated graphics issues, while they're all built in the processor, the igpu could have a failure, thats in fact how those KF/F intel processors came around, a bunch of them came out of the lito process with failure on igpus so they just disabled them.

anyways, i dont see many options there, considering it doesnt crash without drivers or safe cuz both would use the integrated graphics with only basic functions so that you can actually have video and can install the gaphics, no wonder it doesnt crash then, to me, everything suggest when you igpu is called to render stuff it fails at some point, resulting in your video crash

what i would do? get a test dedicated GPU and test, if everything goes fine as heck, i'd say RMA the chip... or go dedicated gpu and disable IGPU
 
Feb 8, 2020
9
0
10
anyways, i dont see many options there, considering it doesnt crash without drivers or safe cuz both would use the integrated graphics with only basic functions so that you can actually have video and can install the gaphics, no wonder it doesnt crash then, to me, everything suggest when you igpu is called to render stuff it fails at some point, resulting in your video crash.

Um, I ran furmark gpu test for 1 hour with no bsod, it was at 99% gpu core usage. This wouldn't have happened if the gpu cores inside the processor were faulty. Also, the display comes back for a sec if I press win+ctrl+b+shift as mentioned in the post.

I can't get my hands on a decent dedicated gpu (I'm just 16) and that's why I bought an apu, I just wanted to run some stuff like crysis, and cod 4 because I used to play em when I was 11 :(
 
Um, I ran furmark gpu test for 1 hour with no bsod, it was at 99% gpu core usage. This wouldn't have happened if the gpu cores inside the processor were faulty. Also, the display comes back for a sec if I press win+ctrl+b+shift as mentioned in the post.

I can't get my hands on a decent dedicated gpu (I'm just 16) and that's why I bought an apu, I just wanted to run some stuff like crysis, and cod 4 because I used to play em when I was 11 :(

its all speculations, all that you have said to me only suggest a igpu failure, the washed up screen with background sound still going is a typical issue when, say, you have a dedicated gpu and overclock it too much, the driver crashes the screen gets black but u can hear the computer still going in the background, plus, if you alt tab you can still see other windows, wich leads back to suggest that the d3d or whatever kind of rendering said game is using, crashed, normally reinstalling like mad, as you did, fixes the issue, if it doesn't, well, everything is again speculations and what you said , again, suggest a faulty igpu.......

notice that just because it works on furmark it doesnt mean its not faulty... now what? the only thing you can do in said situations is do a clean driver reinstall, test, if not, go for a whole system reinstall, fresh o.s fresh drivers, still failing? boi, to me, hardware issue, but you already did that... so lets see what others say.
 
but then again, why not explore more just for the sake of it? reset your bios by removing the battery on the motherboard and wait a minute just to be sure, put it back, if you have 2 ram sticks remove one and change the slot and test it like that, then test the other alone, if you already have just 1 stick, change it to another place.
 
Have you enabled XMP settings in UEFI, or whatever MSI refers to these parameters for Ryzen. I'd also recommend Memtest86 to check RAM, as the iGPU uses system memory to function on these APUs. A small bump to memory voltage could help. I had to do exactly this to fix crashes/display corruption on a build I setup for family with a Ryzen 2200G.
 
Solution
Feb 8, 2020
9
0
10
I hav
Have you enabled XMP settings in UEFI, or whatever MSI refers to these parameters for Ryzen. I'd also recommend Memtest86 to check RAM, as the iGPU uses system memory to function on these APUs. A small bump to memory voltage could help. I had to do exactly this to fix crashes/display corruption on a build I setup for family with a Ryzen 2200G.
I have memtested the rams (0 errors), I've mentioned it in the post.
 
OK. Is the latest UEFI/BIOS itself installed? Since OS was installed several times, does the issue occur right after install, or after certain driver or order of drivers are installed? It is also possible the iGPU is faulty as mentioned. Have you tried an extra bump in memory voltage. Some boards have vdroop vs. actual and needs a bit extra to compensate. Saw this myself with ASRock board and vengeance RAM combo.
 
Feb 8, 2020
9
0
10
but then again, why not explore more just for the sake of it? reset your bios by removing the battery on the motherboard and wait a minute just to be sure, put it back, if you have 2 ram sticks remove one and change the slot and test it like that, then test the other alone, if you already have just 1 stick, change it to another place.

Yeah, dude I tried resetting the bios to factory defaults and one ram at a time too.
 
Feb 8, 2020
9
0
10
OK. Is the latest UEFI/BIOS itself installed? Since OS was installed several times, does the issue occur right after install, or after certain driver or order of drivers are installed? It is also possible the iGPU is faulty as mentioned. Have you tried an extra bump in memory voltage. Some boards have vdroop vs. actual and needs a bit extra to compensate. Saw this myself with ASRock board and vengeance RAM combo.
Yes, bios is latest that was released in 2019 and I've experimented increasing the voltages.

If igpu is Removed as you said, how did it pass a gpu burn in test by furmark for an hour, I don't understand :/
 
Last edited by a moderator:
Feb 8, 2020
9
0
10
Your specific kit? 1st Gen Ryzen CPUs and APU's were very picky vs. 2nd Gen (zen+) and now Zen 2. Brands by themselves didn't guarantee compatibility. What about other questions, and updates available on these? Any changes since this all started?
I solved the problem by getting a replacement of the ram, because it was faulty. Not sure why Memtest86 said 0 errors tho. I got the exact same model of ram, now everything is fine. It seems like the other set of ram was sending corrupted data, hence the black screens.

Now, how do I mark this a solved?