PC Crashes In Games & Graphics Artifacts, GPU Temps Seem Normal

[LonexSoul]

Hey guys,

I've been having some issues with what I think is my GPU for a while now, and they got a lot worse a few days ago.

About a year after I got my graphics card in 2013 (see specs below), it started making really bad artifacts about 30-60 minutes into playing a game. They looked very similar to what the person in this video was experiencing. The artifacts would always randomly appear in various forms such as glitches on specific characters, objects and terrain or sometimes they would be lighting and texture issues. They started appearing when my warranty period with XFX ended, so I couldn't RMA the card and had to live with them. They didn't bother me too much because I could eliminate them in some games by changing between windowed and full screen mode or restarting the game, and they would sometimes not appear again.

The real issues started a few days ago when my PC would completely freeze on a grey screen with vertical lines and the audio would loop robotically on the last sound before it froze, at which point I would need to restart my PC. This again only happened while playing games (I've only tried Dota 2 and AoE 3 since it started happening). I was surprised when it happened in AoE 3 since it's not a particularly graphically intensive game and I wouldn't expect my GPU to overheat from it. A handful of other issues started happening too, I've listed them below:

• 
  ■ Some programs would slow my PC to a crawl for ~2 seconds when they start (only noticed this with OBS Studio and AoE 3), this would cause audio to become choppy and robotic, and friends noted that my microphone sounded like that as well.
  ■ When putting the PC to sleep and then waking it up the next day it would sometimes (but not always) go to the same grey screen that would require a restart.
  ■ Sometimes when waking up from sleep, the PC would turn on fine, but any interaction with it would cause the screen to go black for ~2 seconds and then turn back on with a notification popping up saying "Default Radeon WattMan settings have been restored due to unexpected system failure".
  ■ When the PC would go into the grey screen and I restarted it, it would sometimes turn on but the screen remained black and my mouse and keyboard would not turn on. I had to leave the PC off for ~10 minutes and then try again, and then it would turn on just fine.
  ■ Not sure if this is significant, but my GPU's idle temperature was 60 degrees Celsius when I checked.

I find this bizarre since I never saw my PC overheat or the fans run particularly loud and I've been cleaning it from dust fairly regularly, so I struggle to imagine that it is an overheating issue that is causing my problems. But then again I can only imagine this being some problem with the GPU since it only happens in games. From my reading of similar issues, some people also noted that it could be a PSU issue, but the graphics artifacts I've been experiencing make me lean towards this being a GPU issue. Also, my PC went through multiple fresh installations of Windows 7, 8 and 10 over the years and I've been keeping my AMD drivers updated as often as I could and the issues persisted, so I don't think it has anything to do with software. I also never tried to OC my hardware either.

I decided to stress test my GPU with FurMark yesterday and recorded it hoping to capture the moment it would crash and I did. Here is an image with CPU-Z, GPU-Z, and Speedfan running on my second monitor just seconds before the crash. And here is an image of how the screen looked when the crash happened. FurMark ran for about 7 minutes before the crash happened, and I think it was the loudest I've ever heard the GPU's fan go. The temp. never went above 84 degrees.

Right now my PC is fine running Chrome, live streams, HD videos, IDEs and text editors all at the same time, but it crashes 1-10 minutes into a game, any game. I'm planning to do a full system upgrade in about half a year, so I wouldn't want to waste money on a cheap GPU for this time window if possible, but I'd appreciate a confirmation on this being a GPU and/or if there's any workaround that I could use to keep my PC alive for a little longer.


Thanks for the help!


Specs: Speccy Image

• 
  ■ CPU: AMD FX-8350 (w/ Corsair H100 cooler)
  ■ RAM: 16GB DDR3 (2x8GB)
  ■ MB: ASUS M5A99X EVO R2.0
  ■ GPU: XFX AMD HD 7970 DD Edition (Driver version: 17.4.1)
  ■ PSU: Corsair TX750M 750W
  ■ OS: Windows 10 Pro 64-bit

 

[Jonathan Schiller]

I had that problem with my old PNY hd7950. It appears it was caused by heat damage on the card. Unfortunately I didn't catch it in time and had to replace it. I found The Gray lines on my screen would crash with gpu frequency, not temp. I got mine stable for a little wile by underclocking it below the crashing frequency. But with your card crashing on low gpu games like DotA and aoe3, I would say you are past that point. On a good note you can get about the same performance as the 7970 with a new card for under 150 USD like a rx 470 or gtx 1050 ti

 

[dontlistentome]

You could try removing the heatsink/fan and reapplying thermal paste. It's possible that might keep the videocard from heating up enough to trigger the crashes. I do think it's heat since it takes awhile(minutes) to happen after you apply a load. The heat is building up during that time. If the card is permanently damaged then keeping it cool may not help, since a component on the card may be failing and there's not much you can do about that.