Question DWM.exe crashing after Windows Update

Sep 7, 2021
3
0
10
0
Hello All!

I'm running into some difficulty with my computer so I'm going to splurge it all out.

Windows 10 Pro: 21H1
CPU: Intel 8700k
Motherboard: Asus Maximus Hero X
RAM: G-Skill 32GB
SSD: 1TB NVME M.2 Samsung Evo 960
PSU: Corsair 750W Gold
GPU: GTX 960 / Asus TUF OC RTX 3090
Monitor 1: (Display Cable) Dell 4k 60Hz
Monitor 2: (Display Cable) Dell 2k 144Hz G-Sync

(I know the specs could be more accurate but those are the pertinent)

Ok so here's what's happening:

I initially had the Asus 3090 in the computer but after the Windows update to 21H1 when running Overwatch it would randomly have both screen shut off showing no input but you could hear the sound in the background. First in order to replicate the issue, I downloaded FurMark and ran a stress test so I wouldn't have to wait for the issue to reoccur. Upon starting the stress test the computer immediately showed the problem again. While it was in this state I was able to log into the computer using TeamViewer and view what WOULD have been on the screen through there. There were no errors present but curiously enough I noticed in the Device Manager there was no more display adapter listed, not even a standard Microsoft one.

Step 2 was to see if my GPU was disconnecting. To do this I continued the stress test and noticed that the GPU had almost a burning smell coming from it. I know that the smell was coming from it so I decided to RMA the card back to Asus and put in my 960 hoping that would fix the problem. And for a couple hours it did. Then the problem came back. Same problem, same symptoms. I took a look at the Windows Event Viewer and Windows Reliability which showed a bunch of warnings for USB drivers and such which happens but the main one which caught my eye was, "Windows Desktop Manager" had an APPCRASH with a, "KERNELBASE.dll" associated with it.

I caught advice from Dr.Google and found that many people have had the same issue but it isn't correlated to the windows update. Furthermore, nobody posted a solution to the issue at all. I ran a SFC scan and a DISM scan, both which were fine. Did the DDU and reinstalled drivers, no success. Updated the BIOS, reset the CMOS, tested RAM, tested PSU, etc, etc, you name it. I'm about to start a support group for this issue because its driving me crazy.

FINALLY Step 3 today was to reinstall windows on a second drive. Hallelujah there was no issue I was able to run the stress test with no issue! I went back to my original drive and did a window, "Restart." Essentially that downloads windows from the cloud, saves all your files, wipes out all your apps, and reinstalls a fresh copy of windows. I ran the stress test and NO issue again. Great, let's install drivers now. Installed Geforce Experience and the latest Game ready drivers and immediate stress test fail. So I think I've got my culprit. I uninstalled everything with DDU then let windows install the drivers. Stress test works fine! I closed the stress test and on a whim opened it up again to make sure it was ok and IT FAILED AGAIN!

Now, Im pretty handy with computers and would love to post a MiniDump for you, which I don't have because its not BSODing. I'm stuck here with DWM.exe APPCRASH driving me nuts and no idea where to go from here. Its almost completely arbitrary. I'm looking for some help with a technical fix, not a, "back up your computer, reinstall windows, and never update it or make eye contact with the logo," kind of fix.

Any fresh ideas are welcome and appreciated, I promise if I can find a fix I will report what the solution was!

-Justin
 

Colif

Win 10 Master
Moderator
if its not a BSOD, its not software.

Windows Desktop Manager sits between Applications and the Graphics card drivers

Have you tried just running off the integrated graphics in the CPU to see if it too causes problems, as it is a 3rd gpu to use and it doesn't cost you anything
 
Sep 7, 2021
3
0
10
0
if its not a BSOD, its not software.

Windows Desktop Manager sits between Applications and the Graphics card drivers

Have you tried just running off the integrated graphics in the CPU to see if it too causes problems, as it is a 3rd gpu to use and it doesn't cost you anything
I just tried that today and so far it hasn't had an issue. I'm going to test it after an hour of use and see if it continues. I had a thought also of it maybe being the cord going to the GPU from the PSU. That OR maybe a capacitor in the PSU for that channel is going out which is causing an intermittent issue where it interrupts power to one line of the power cord to the GPU. This might account for why the dwm.exe is unable to recover?

Idk, I'm shooting in the dark here.
 
Sep 7, 2021
3
0
10
0
Ok keeping everyone up to date on this. I noticed that the issue only occurs occurs when the video outs were plugged into the GPU. The fact that it was happening to both GPUs led me to believe that it was either in front of the GPU or behind it. TURNS OUT that it was my 8 pin 12V cord that was plugged into my PSU. one of the wires on it had burned back and had some green oxidation on it. Also 3 of the 8 pins had melted into the port. I'm going to RMA the whole PSU since its a Platinum not a Gold and has a 10 year warranty. I plugged a new cord in and changed it to a different plug on the PSU and I haven't had an issue yet! My theory is that my 3090 drew too much power over a long period of time with Mining and it degraded the sheathing on the cord (likely due to manufacturer defect or freak accident) After I stopped mining on it, the exposed wire was causing the plug head to heat up and the wire was degrading. This resulted in inconsistent power to the GPU. When I would run a stress test, the GPU would request a surge of power to ramp up and it would immediately short and drop in voltage. The drop in voltage threw an error to DWM.exe because it couldn't provide the power. As a result it couldn't recover.

Anyone have any other thoughts?
 

ASK THE COMMUNITY

TRENDING THREADS