I have been having a lot of issues with my graphics card when playing games including the drivers crash and kick me out of the game, everything completely freezes and I have to shut down the computer, and the computer just turns off and restarts. All of these seem to happen completely randomly.
- Intel Core i7-2600k
- AMD RX570
- Kingston KHX1600C9D3/4GX 9905403-198.A00LF KHX1600C9D3/4GX 9905403-198.A00LF16GB Ram
- Asus P8P67 REV 3.1 motherboard
- Evga 500b power supply
Things I have tried:
- Reinstall graphics card drivers
- Turn off fast startup
- Update windows and drivers
- Turned off freesync
- Increase voltage in tuning
- Increase power tuning
- set all AMD driver settings to factory default (this is what I have right now so the tuning stuff isn't active)
- used a DDU to uninstall drivers then reinstalled
- reapplying thermal paste to GPU (might have helped a bit because they seem to happen less often
- uninstalling and reinstalling audio drivers
- removing all duplicate hidden devices in device manager.
Games like Terraria and Minecraft seem to be unaffected and after doing everything above Assasin's creed origins seem to run fine as well. GTA5, *Warframe, Just Cause 4, Hitman 2, and Rainbow Six Siege all still have problems though.
Using CPUID HWMONITOR and AMD's built-in monitor's voltage and temperature doesn't look like it affects the issues.
*Got Warframe fixed by verifying the cache every time I play.
Solved! Go to Solution.
After trying a bunch of different combinations of cables and adapters without any luck I'm just going to return it. I've put an old 1050 ti in instead and literally every single problem I've been having is gone. It's kind of an unsatisfying answer but I guess this is what warranties are for.
Ok, I found this thread (https://linustechtips.com/main/topic/898055-black-screen-gpu-fans-at-full-speed-crash-fixed/) where the guy was having the same problem as me. I looked at my GPU and replaced the thermal paste. So far I've had no freezing or black screen crashes though it's hard to tell because I could just be on a lucky streak. Still having problems with my AMD drivers crashing though, I might have to wait till an update comes out to fix that part.
Ok just as an update I'm fairly sure I found the problem. Using these sites Power Supply Calculator – PSU Wattage Calculator | Newegg, Power Supply Calculator - PSU Calculator | OuterVision, I found that I'm pushing my power supply right to its max (Rx 570 uses like 100 more watts then my old GTX 960) so I'm gonna order a new power supply and hope it works. It might take a while to get here because of certain viruses.
One more way you can do is decreasing the limit power, down to -20% ~ -10%, and increase the fan speed.
It worked for me. No need to change the PSU, but the performance will be affected a little bit. Keep the temperature between 60-70 *C at the full-load is the best way to stable your GPU.
I found out that when I'm at the full-load and when I do nothing (idle), the fan speed did not change too much (about 1200 rpm). So I have to edit the fan curve in the Radeon software tunning.
Changing the power limit does not decrease required voltage. It only changes the power ceiling that is dynamic and only used when needed anyway. Many cards that come with factory overclocks need a raised power limit to be stable. So you would need to lower the actual voltage more than the power limit to use less power from the power supply. Chips however are designed to operate between power levels. Just like overpowering them, under powering the chips can reduce chip life too. As the OP pointed out there are other cards at that performance level that run on less power. Both from green team and in the Navi lineup.
The temperature as you point out is very important to the stability of the Polaris chips. While they claim the limit is 95, my experience is that they start to become very unstable around 74.