Drivers & Software

[davejb]

TL/DR: see 2nd post below. Power supply was the problem.

This is the error I'd get: "Display driver amdkmdap stopped responding and has successfully recovered."

Months ago, when I first upgraded my RX 280 to a 580, I was sad, mad, furious etc when I had just horrific stability problems. I may have finally isolated the causes. I'll just paste my first post then today's update.

NOTE: my PC specs: Vega RX56, Intel i5 cpu, 16 GB RAM, Windows 10 1709, I believe I'm on the AMD driver release about 6 to 8 weeks ago (18.1.1?).

Older post (maybe 5 months ago?):

I wish I had to time to put up a thorough post on this, but the short version is that about 3 months ago I switched from a very stable (for about 3 years, IIRC) gaming platform with an R9 280 to an RX580 (didn't change any other part of the PC). Once I switched to the RX580 I had constant crashing. Like, just sitting at the desktop.

It's still not perfect but I started taking notes and isolating one thing at a time and I found that I picked up a ton of stability by using in-game settings for anti-aliasing and anisotropic filtering, that is, leaving the driver settings at default (use application settings). This is for NBA 2k17 and Mass Effect Andromeda. I then made sure any adjustments to visual settings were done in game.

I know that's not super thorough but I hope that helps. You might do a full uninstall with the AMD cleanup utility and DDU and change only one setting at a time for a while. I can tell you I had crashing so often for a while I nearly took a sledge hammer to my (expensive!) computer and monitor.

Today's new info:

One more update. I believe my issues starting with the 580 through my current Vega RX56 can be narrowed down to two things: incompatible driver/game video settings (this has been sorted out for a while) and Windows 10 updates. Recently after having about 5 or 6 weeks of stability, I started having the same issues as before. That is, the "Radeon settings host application stopped working" error both while playing and sitting at the desktop and the error that says your video card might have been removed. Seriously, I'd get the Radeon crash often while just sitting at the desktop.

It got very bad again just the other day (Monday 3-19)after weeks of stability and I realized the only thing that had changed is the March Win 10 Cumulative update. That is, I had stability for weeks then one night after I'm done gaming, run updates and shut down. Come back a few days later and constant crashing again. I uninstalled the updates, rebooted (then accidentally uninstalled Feb and Jan, which may have been needed).... in the end, NO OTHER CHANGES and right away back to super stable.

Hope this helps you guys. I know there were some

[davejb]

Well, curious. I wonder if some of those changes might have softened possible transient power spikes? I think I've read the Vega (and maybe 580) have bigger transient spikes than in the past? Anyway despite all the time that has passed since this post and all the things that seemed to help sometimes even for weeks, recently I had another HORRIBLY maddening run of random crashes in game.

PROBLEM SEEMS TO BE SOLVED: power supply. God bless it.

So, I had quality Antec 750 watt power supply that I think I bought in about 2011. Right up until I replaced my R9 280 (first with a 580 then Vega RX56) I had incredible stability so it never occurred to me the PS could be the problem. I'm pretty sure the Vega RX 56 doesn't draw MORE power than an R9 280 but I have read that possibly it does spike quite high sometimes.

Anyway, I bought a brand new Corsair HX850i and have now gamed for at least 12 hours with no crashes. Interestingly, the power drawn through this PS shows that it's not pulling even half of 850 watts so it's total overkill apparently. I mean, the power supply fan hasn't even needed to come on yet.

[davejb · Accepted solution]

Going to just paste in here what I posted in this long thread. This is the original thread I've been working in:

https://community.amd.com/t5/graphics/display-driver-amdkmdap-stopped-responding-and-has-successfull...

Here's my newest post about this. Hope it helps:

SON OF A !!!!! YEARS LATER, I figured it out. It's fixed. NO MORE CRASHING.

So, I've mentioned in this thread and others that I had none of these issues until I replaced my older (RX580?) card with a RX Vega 56. I'm not sure of the timing exactly but somewhere around that time (or maybe exactly the same time) I also replaced my monitor with a VIOTEK GN32Q monitor because it supported Freesync. It turns out that it's somehow related to the lack of a monitor driver (with a proper list of supported resolutions and refresh rates, and possibly FreeSync range) for this generic monitor. Bear with me as I'm not at home on my gaming system to have to guess at the names of some of the thing below.

I stumbled onto this thread here. Like, him, the AMD driver showed my monitor's supported Freesync range at 90 to 144 Hz. FWIW, the game I really struggled with the crashing in is F1 2020 (Codemasters) and I was running Anti-lag, Enhanced Sync and FreeSync. My in car driving frame rates are, IIRC, in the 80's. I think the crashing was caused when the game would cross the 90 FPS range and then maybe the driver would try to switch to FreeSync.

I used the CRU utility to change the FreeSync range from "90 to 146Hz" to "40 to 146Hz". I also edited the monitor's Detailed resolutions to have only two settings supported, the 2k res at 60Hz and at 144Hz. This took a LOT of experimenting and reboots but in the end it's FIXED for me. I've even turned back on the TDR setting with no ill effects.

Changing this stuff isn't perfect. When sitting at the desktop (windows 10) I sometimes have to use the display settings to switch back and forth between the refresh rate, between 60 and 144 Hz. For the most part, once I'm done gaming I change it back to 60Hz.

I'm not sure how useful it would be for me to walk anyone through every step since you may not have the same monitors. This is a lesson for me to never buy a generic monitor again. This monitor, I love everything about it and it cost me only $450 while supporting FreeSync and the other certified FreeSync monitors at the time were starting at $900, IIRC. I am referring to monitors with a minimum size of 32" BTW.

I hope this helps get some of you on the right track. I can tell you that many nights, for the first 30 minutes my F1 game would hard lock (I had TDR turned off) sometimes 5 or more times then for whatever reason would usually settle in for the next few hours. Now, I've had at least 5 gaming sessions with not one crash of any kind.

Poor AMD, this was never even their fault and I can now see why while this problem wasn't that rare, there were many people having no issues like this. I'll criticize AMD in one way, they should have a good bug reporting tool that could have tracked that many of us have generic, uncertified (in terms of FreeSync) monitors.

Lastly, I'll help when I can but I'm super busy and it may be days until I can respond here. Suffice it to say, you need to understand how to use CRU, you need to know your monitor's supported refresh rates, and you need to have big cajones because you may end up having display issues if you mess up the CRU settings.

Oh, one more huge frustration I had with F1 2020. Until I figured out the problem, the game would sometimes be blurry and not smooth at all. I'd look in the games video settings and saw F1 had switched to a customer resolution that was just below the native 2k resolution (3840 x 2160 or something?). I couldn't fix this no matter what. It turns out that the per game AMD driver settings must not be on anything but Full Panel. Maintain aspect ratio was the setting that was causing this. It seemed like the right setting but in hindsight was not.