Display driver amdkmdag stopped responding and has successfully recovered.

Omg this actually worked. Thank you so much. Been struggling with this for MONTHS. Cannot believe it was something so simple.

No luck here. Tried with everything said in the post + reducing the refresh rate, but it still crashes. Funny thing: VRAM clock does go down to idle levels but only when I have one monitor connected.

Feels terrible to know that, due to either Microsoft's or AMD's lack of action (or both, most likely) I'm either restricted to not play certain games or not use my entire setup.

Yes I think the drivers are a little rough/buggy at times. For most people who have it for prolonged periods I think it's bad RAM, RAM timings, or a failing GPU.

In my case it was the driver bug that causes max VRAM clock on certain refresh rate/screen combinations. I have not had a crash since using 142Hz. Interestingly in my case though the issue is much less frequent with single channel memory at 144Hz which is very annoying as you've mentioned.

Oh, and I tested my GPU gaming in Fedora Linux at 144Hz and it never crashes which really means it's a Windows/AMD Driver issue! (in my specific case with the max memory clock at 144Hz bug)

This issue amazes me... over the last 5 years I have tried million different things, but it still persists and by the looks of it still to this day nobody has figured out the definitive reason why it happens.

So I read all 16 pages of the thread and it seems all the same suggestions and potential causes are being discussed:

* undervolt/overcolt the GPU

* increase/decrease power limit

* downclock GPU

* uninstall additional software (like Radeon software i.e. ReLive)

* uninstall High Definition Audio

* issues linked to 120/144Hz monitors or DisplayPort

* issues linked to PSU

* one solution which I have never seen before was related to dual/single-channel memory, but it seems that has not definitively fixed the issue.

Have I missed anything else?

Just a little bit of my story - I had this issue probably for 5 years if not more. I had it on R9 290X, R9 390X, R9 Fury X and on RX5700XT. All except RX5700XT were purchased used. At the same time, I only had a single Nvidia card and that was GTX260 and it did not have this issue. I always assumed this is AMD issue, but I can see some people reporting it on Nvidia as well. Now if there is a suggestion on the internet on how to fix this - I bet you I have tried it.

I always had an overkill power supply (Enermax Platimax 1500W) from the days of me owning 2X4870X2 on quad-CF, later on I have downscaled to a more "reasonable" Corsair AX1200i and AX860i, so that is not power supply issues either.

What makes this issue particularly annoying is that the causes for it are always different, for one person it will crash in Warzone, but not in Whicher, for another person is exactly the opposite, for one person the cause is HDAD, for others it does not.

Now there is an interesting twist on my story... my RX5700XT was a new card with a warranty and it turned out to be defective in more than one way. Eventually it was RMA'ed and thoroughly tested by the manufacturer before it was sent back to me. Guess what - it doesn't have this issue. And it wasn't 2 days in a row, or 2 weeks. I got RMA back in May and since then I have had no issues with driver stopping.

So my conclusion here - AMD drivers are kind of ****, they have been unreliable for as long as I can remember, but importantly for this issue there are:

* your card is defective, even if it is rejected by RMA (they just trying to stonewall you).

* the way windows handles the video drivers is unstable and even the smallest issue will trigger the driver crash - that is why so many different reasons exist... literally anything can cause the display driver to crash and it is just not acceptable.

What is the solution - buy a new graphics card thoroughly test it before 30 days the return policy expires, if you get a driver not responding error just return it and do it as many times as it requires. If your GPU does it then it is faulty, and even if it is not faulty you still don't want to get involved with this issue... as there are quite clearly no solutions for it. Windows are just **** like that and unstable (not like we have the option to use anything else), so to avoid the issues you need a perfectly working GPU.