Display driver amdkmdag stopped responding and has successfully recovered.

Driver 21.8.2 fixed it for me, just update like normal (no need for Factory Reset or DDU)

Updates, people, updates.

Yeah I think there's a link to hardware, specifically the memory chips used on the GPU in some cases and the monitor timing (monitr VBLANK is what I think it's called).

For example, I have a 5600 XT. For some reason the VRAM gets locked to 1750Mz at certain refresh rates. This is OK for some cards, but cards like mine can crash. Some RX 5600XT models can be flashed to effectively become an RX 5700, but it depends on the memory modules. I'm sure there's some link between drivers, memory modules, refresh rates, and the phase of the moon for all we know.

What's real interesting is I can overclock my card VRAM above 1800Mhz no problem once the refresh rate is set to 142Hz or less. So it's not just VRAM speed that changes, but also some subtimings similar to regular RAM I bet.

I wrote a blogpost about my personal experience and fix. It's a shame that it's not a universal fix, but some of the issues are definitely linked to failing cards and poor drivers: https://evanshortiss.com/amd-driver-memory-lock-timeout

I gave up. I did a new build. New case, new motherboard, new cpu, new cooler, new ssd. I kept some components from my old build (gpu, psu, ram) BUT I have already switched out these components at one point or another on my other build. Nothing here is from the original build except 1 stick of non-faulty ram and I have had tdr failures through every build, including this new one which was quite costly and I thought would surely fix my problems. I have set this computer up in multiple households which rules out any problems related to the power going into the computer.

I have been EXTREMELY careful and precise when building this pc and when troubleshooting in windows. I’ve done several windows installs from scratch. I’ve spent a lot of time troubleshooting, installing only the minimum software needed to test.

Nothing is overheating, my temps are fine.

No graphics driver installs or uninstalls of various versions have changed anything.

None of the non-conventional fixes from users in this thread have worked for me.

No stress testing of any kind (gpu, cpu, and ram) has created any errors.

No amount of gpu underclocking or overclocking changes anything.

No ram configuration, refresh rate, or minimal usb usage has any effect.

I thought maybe I’ve had super bad luck and both my old gpu and current gpu are both faulty, but my girlfriend has my old gpu in her pc and it works fine, so I am convinced neither gpu is the problem.

i don’t know what to do. How can I build a whole new computer, do a minimal windows install, and still have the same problem?

Someone please help, I am embarrassingly desperate.

Hello! The problem was that the graphics card had gone bad, so it worked in basic mode without acceleration (BIOS and simple 2D graphics with Microsoft Basic Display Adapter), but when I installed the drivers it failed to a black screen. It was out of Warranty, so I decided to try fixing it myself. I removed the fan and heatsink, cleaned the thermal paste from the AMD chip and then used a heat gun at about 260 degrees Celsius for a few seconds to airflow repair the card. I heated mostly the AMD chip, but also the memory chips and the rest of the card. I then put fresh thermal paste and put back the heatsink and fan and booted into Windows again. It worked! Now I can install latest AMD drivers properly without any glitch. In some performance test maximum temperature was 65 degrees Celsius without any issue. I'll mark the thread solved. Thank you all.