AMD Radeon 5700xt
Windows 10 64bit
Radeon Adrenaline V20.7.2
LG Ultrawide (21:9) with HDMI connection
Gigabyte X570 UD Motherboard, Rev. 1.1
AMD Ryzen 5 2600
Corsair 850W RM850x PSU
16GB of RAM
So i built this computer from scratch about 6 months ago, and about a week after i first started with it i began having constant crashes from certain games, with Stellaris, Arma 3 and Dying Light being the 3 biggest ones. Nothing in particular will trigger this crash, it could be 3 minutes after i start a game, could be once an hour or i could go 5 days smoothly and the issue starts occurring again. Stresstesting, benchmarking, and PC health checks don't raise any red flags, all (ALL) of my drivers are up to date. I've looked at every online tutorial and even a few locked threads on this forum and followed the fixes they suggested, without any success.
This is my last resort, or I'll just have to call on the warranty and swap out for something less buggy like an Nvidia card because this is seriously game breaking.
The problem may also be in the bios, updating the bios can help, but this is not certain.
Also try playing with the GPU Scaling on / off (try this first), if you play with off.
"Black Screen of Death" - you mean BSOD or black screens?
Try disable all graphics enhancements in GPU driver and if it happen again, then try figure out what is exactly same every time. When I bought RX 5700 I had same problem with Witcher 3, but I and others we reported it and now everything is fine (problem was with shadows set above medium) - it was fixed in February driver release. Not so crucial problem was also with DeusEx Mankind Divided, where I noticed application crash near main squares (in fact it was caused by metro nearby) - it was fixed in July driver release. Reason why I am telling you all this, is that AMD repairs problems, but they need raw data = savegames, exact description, etc. With general descripton is hard to find a bug.
Nothing sets it off. As i said, i can play for hours without an issue, then i'll crash 23 times in a day. Tried disabling all the fancy graphical fluff, still happens.
I can't make this happen on demand, it is entirely random.
I had this issue. Did TONS of tests. My PC would hard crash giving me a motherboard debug code of 00. I at first thought it was my CPU, but after replacing the cpu the crashes continued. When I say tons of testing.. I mean the most testing I've ever done to resolve an issue with my PC. A week of nonstop troubleshooting led to me a "workaround" rather than a fix. I was able to boil it down and recreate the crash when running my GPU at over 150watts. If my GPU spiked or stayed above 150watts my PC would eventually crash. I limited my power slider within Wattman to a -33% to keep my GPU watt usage under 150. After doing this I have not experienced these crashes again. To clarify I have the Sapphire 5700xt Nitro+ card. While limiting the gpu to just 150watts it still maintains 1770 - 2000mhz on the core (depending on the game and scene). So while I'm not completely happy with this I still get great performance while staying cool and quiet. I don't know why going over 150w on my GPU causes the crashes (and its every game). Every driver update I do I check to see if it's fixed or not. Hope this helps you.
Keep an eye on your GPU junction temps and your GPU Core clock speeds.
My Sapphire Pulse RX 5700 XT loves to push itself to 2029MHz if I leaved it on auto. After a while of this my games crash. Sometimes I get a black screen sometimes it goes back to desktop. If I check event viewer it says that the driver stopped responding but recovered.
The Fix: I limited my GPU to 1900MHz and 1000mV in AMD Radeon Software. It seems my GPU is binned pretty decently for thermals but it struggles when it's left to freely push itself.
These are my GPU's clock speeds
My GPU would stay at 2029MHz constantly while gaming. I was instructed to try this by someone over on Reddit so I figured I'd share it here too. Check your GPU specs with the manufacturer and see if your card is exceeding their specs. If it is try limiting them to see if it solves the driver crashes for you guys.
This has essentially eliminated all driver crashes I've had except for Chrome/Edge black screens which apparently has something to do with AMD drivers and chrome using a hardware overlay for video playback if I'm understanding correctly.
It seems to me that some GPUs simply have a hardware defect that appears from time to time and it is unlikely that a driver will be released that can fix it. If you have tried all the diagnostic methods and are sure that your PC is working properly (for example, it works with another gpu without issues), use your warranty. I spent almost a year looking for a problem in my pc, endlessly checking memory and cpu for errors, changing power supplies, but the problem did not go away. Now with a different gpu my pc is rock solid, and tomorrow I will RMA my rx5700
So after some fiddling around, i've mostly mitigated the issue by enabling power tuning and setting the "Power Limit (%)" slider as far to the right as it goes. It's been a week and i've only had one crash as opposed to several a day. The issue is definitely related to some kind of GPU spike.