I'm using a VisionTek RX 480 8GB GPU, which has been working great for four years. But now it seems to be causing issues where the driver will crash at odd moments. It doesn't appear to be related to heat - I've seen it at less than 60 degrees moments before it crashes, and it will occasionally crash on boot.
The presentation is usually the same: the video will freeze for about ten seconds, and then the display will shut off. Unplugging and re-plugging the HDMI out doesn't help, and the computer won't respond to any input, including signals to shut down. I have to hard power off each time, even after leaving it in this state overnight. Windows Reliability Monitor will register a "Hardware Error" when this happens.
The cause is difficult to pinpoint. If there's any pattern, it seems to be slightly more likely shortly after exiting a full screen game of some sort. But it can also happen when just using Firefox, or a few times, when it's just sitting idle.
I've reinstalled three different sets of drivers, including Factory Reset installs and using DDU to completely remove all traces of the driver beforehand. No manual overclocks or tweaks applied. The issue still persists.
The rest of the system is:
Anything I can do to help diagnose?
PSUs can start to output less power as they age, not sure if that is your problem but it is good knowledge. If you can test with a known working system or try a different PSU that can help you isolate the problem.
What was the PSU Output during the tests for 12 & 5.0 & 3.3 Vdc? They all should have been -/+ 5% tolerances. For instance the 12 Vdc should have shown a minimum of 11.4 Vdc or higher during the testing.
Run OCCT GPU MEMTEST and see if your Ram check good without any errors.
Here are some GPU testing software from this tech site: https://beebom.com/best-tools-stress-test-gpu/
Also run MEMTEST86 or Windows Memory Diagnostic just to make sure your System RAM Memory is physically good and not going bad.
If the PC is crashing using a Stress test, at least, that indicate you have some sort of issue with either the GPU card or driver or some other hardware. I hope you get it solved.
Thank you. I'm trying OCCT as I type. In the meantime though, I'm afraid I don't understand what "PSU Output during the tests for 12 & 5.0 & 3.3 Vdc" means. Is this a software diagnostic of some kind or do I need a voltmeter or other diagnostic hardware? I don't own any currently and unfortunately don't own another system to test the GPU with.
A voltmeter would be more accurate, but you can use software like Hardware Info 64 to give you the values, and you are looking for it to be outside 5% high or low, 12V should be 11.4 to 12.6, 5V should be 4.75 to 5.25, etc.
If you can, try the card in another system, or try another power supply in the system it's in now to help isolate your problem. If you have a computer shop you could have them do a quick diagnostic on the card, usually most places will charge you a minimal amount, all they have to do is install it and load drivers, a few quick tests.
450Watt seems just too low, and hardware age too.
Have you got much coil whine on the PSU during bench or game?
Did you added/changed any hardware since it was stable?
Can you try another PSU?
Do you OC the card in anyway?
Other stuff you can check.
Testing with basic driver
Re-Seat every component
BTW: I had a 650Watt bronze but my Strix RX480 did not like it one bit, and it sure helps to have a better PSU as my Strix card sure OC's like a charm. 1400mhz super good, plus FSR and still plays Gr8 today!
I agree with @johnnyenglish. You can use a wattage calculator found online and see what your system will draw. The AMD requirement is actually 500 Watts, which you actually want a little headroom so that if you are using a lot of devices, it won't exceed. If you are overclocking you may be drawing more that the 450.