Open World Games freeze

[Sveini101]

Yesterday I was playing The WItcher 3 and after playing it for about 30 minutes my main monitor froze and blue squares appeared all over the screen (on both monitors). The same thing happened in Guild Wars 2, but after five minutes instead. It also happened the minute I entered WoW. It doesn't happen when I play CSGO or hearthstone, so I guess it has something to do with open world games. Maybe its because they require more ram? I don't know. 

The GPU status in device manager says it's working properly, I've tried tried setting the power management mode to prefer maximum performance on the games, I've reinstalled the graphics driver with a clean install, I've installed all important updates in Windows Update, I've tried installing an older version of the graphics driver with a clean install, I've taken the video card stability test and the same thing happened after a couple of seconds.

 

Do you know a way to fix this?

 

PC specs:

- Gainward Geforce GTX 770

-Intel Core i7 4770 3.40GHz

-16,GB RAM

-Corsair 600W PSU

 

Thanks for any help.

 

[Syntaxvgm]

I have a feeling that 

A. You have CSGO frame rate limited and it does not use the gpu at 100%

B. The open world games are using more vram and that is the culprit, but I think more A. 

[AlwaysFSX]

Go do a core burn test and a memory burn test in something like Kombustor or FurMark and see if you get artifacting there.

[Syntaxvgm]

Go do a core burn test and a memory burn test in something like Kombustor or FurMark and see if you get artifacting there.

idk if furmark does it but OCCT pretty reliably will tell you if there's a problem with it. 

[DimasRMDO]

Did you overclock your GPU? If yes downclock it or reset everything to default.

Or the GPU is dying, to be sure do stress test.

[Sveini101]

I have a feeling that 

A. You have CSGO frame rate limited and it does not use the gpu at 100%

B. The open world games are using more vram and that is the culprit, but I think more A. 

Yes the frame rate is limited at csgo, i think.

 

Did you overclock your GPU? If yes downclock it or reset everything to default.

Or the GPU is dying, to be sure do stress test.

No, I haven't overclocked my GPU.

 

idk if furmark does it but OCCT pretty reliably will tell you if there's a problem with it. 

I can't run it in infinite mode so should I use the automatic mode? For how long?

[DimasRMDO]

No, I haven't overclocked my GPU.

Try to set everything down to low, if it's fixed then try to crank it up again until problem appears again. Then go back to the "Safe" settings.

[Sveini101]

When I ran the furmark GPU test the screen froze again. Even when I had low-resolution. And the same thing with the OCCT test.

[Sveini101]

I tried to underclock it by 100MHz, but it didnt work. I've ran the Furmark and OCCT tests a couple of times now, and once the GPU temp got up to 81 degrees celsius.

[Sveini101]

https://www.dropbox.com/sh/rx54iexbj09gu6z/AADW5egpl5mlnIb_MR8oIippa?dl=0

 

Here's the OCCT test pics. The CPU temperatures where at 95 celsius when the OCCT stopped the test. And the GPU was at 82.

[ApolloX75]

https://www.dropbox.com/sh/rx54iexbj09gu6z/AADW5egpl5mlnIb_MR8oIippa?dl=0

Here's the OCCT test pics. The CPU temperatures where at 95 celsius when the OCCT stopped the test. And the GPU was at 82.

Those temps are monstrous. I take it your 4770 has the stock Intel cooler?

Might be time for something a bit beefier if your case has poor ventilation. Maybe even just new thermal paste.

Your GPU could probably use a cleaning too but it sounds like its goose is cooked already.

[Syntaxvgm]

 

https://www.dropbox.com/sh/rx54iexbj09gu6z/AADW5egpl5mlnIb_MR8oIippa?dl=0

 

Here's the OCCT test pics. The CPU temperatures where at 95 celsius when the OCCT stopped the test. And the GPU was at 82.

 

There's an option for the GPU test, error check. Check that box and run it for 10 minutes or so. 

Sorry, I should have been clear about that. 

Then tell me if the test shows any errors on screen before you cancel it. 

[Ryukrieger]

 

https://www.dropbox.com/sh/rx54iexbj09gu6z/AADW5egpl5mlnIb_MR8oIippa?dl=0

 

Here's the OCCT test pics. The CPU temperatures where at 95 celsius when the OCCT stopped the test. And the GPU was at 82.

 

That temperature is a tad bit high, if the error check suggested by Syntaxvgm doesn't return much, I suggest checking your coolings, are the fans gathering too much dust? Fans dying? Coolers inadequate? Thermal paste dying? When was the last time you clean/check your computer?

I've seen this happening due to overclocking (not in your case) and heat (also faulty cards can cause it).

[Sveini101]

Those temps are monstrous. I take it your 4770 has the stock Intel cooler?

Might be time for something a bit beefier if your case has poor ventilation. Maybe even just new thermal paste.

Your GPU could probably use a cleaning too but it sounds like its goose is cooked already.

Yes, I do have the stock intel cooler. I have three fans, and the one at the back has started to act weird, sometimes it starts going faster for no reason.

 

That temperature is a tad bit high, if the error check suggested by Syntaxvgm doesn't return much, I suggest checking your coolings, are the fans gathering too much dust? Fans dying? Coolers inadequate? Thermal paste dying? When was the last time you clean/check your computer?

I've seen this happening due to overclocking (not in your case) and heat (also faulty cards can cause it).

 

I cleaned the CPU cooler and the GPU fans yesterday with some q-tips, I also cleaned the bottom of my cabinet with an old sock. But except from that, no. Fans dying? Maybe^. When I took the CPU cooler of the termal paste seemed ok, but I'm going replace it anyway. 

 

I do have an Corsair CX600 PSU and I read on other threads that the CX600 can damage the graphics card/ other components. Some said that it would only last for 1-2 years.

[Sveini101]

There's an option for the GPU test, error check. Check that box and run it for 10 minutes or so. 

Sorry, I should have been clear about that. 

Then tell me if the test shows any errors on screen before you cancel it. 

 

I did two tests one with DirectX11 and one with DirectX9. The DirectX11 one was a bit wierd, don't think it worked, the log pics doesn't show anything except for the CPU usage, frequencies and memory usage. It says user cancelled after 1min and 2 secs. But the DirectX9 one worked. Sky-high temps as usual.

Log pics Directx9: https://www.dropbox.com/sh/ozl2fitkmn1qwi1/AADGANKiwXjoyFFdPRNhsspra?dl=0