Getting artifacts but not very often.

[JWildOnes]

Im getting artifacts but not often, sometimes it can be none a day, once a day, twice a day and sometimes none for a few days. I took a photo of when it happened not sure if that will help to identify which could be causing it.

My specs are:
MSI gaming 5 Z97
I7 4790k
GTX 970
8gb ram
p2 EVGA 750W
250 GB SSD
2x 1TB HDD

 

[panathas]

Are you overclocking the GPU? If the answer is yes, return it to it's default clocks. It's probably the GPU but before you get there you have to:

1) Try a different cable.
2) Try a different monitor.

If nothing changes then it's the GPU and its VRAM is probably failing. Monitor the GPU temps, make sure that is fans are working/spinning and the card isn't overheating. Since it's probably out of warranty and you can't return it, you could try to lower the GPU and VRAM frequencies. You may be able to reach stability levels in lower frequencies. Just try to drop each clock in small steps (50 MHz) each time (GPU & VRAM) and monitor the results. First try to decrease the GPU clock and then the VRAM. Don't lower both at the same time (-50 GPU test, -100 GPU test, -150 GPU test..., then do the same with the VRAM). That way you will find whether the GPU is failing or its VRAM or both. If it doesn't work after a 300-400 MHz clock drop then your graphics card is dying and you have to get a new one. Now if you do manage to stop the artifacts, keep in mind that this isn't a permanent fix and they may return in the future. In that case you have to try to drop the clocks even further (decreasing the GPU's performance) or get a new graphics card. Good luck.

 

[cryoburner]

That first screenshot is clearly of augmented-reality minecraft. : P

https://images.techhive.com/images/article/2015/01/microsoft_windows_holographic_3d_minecraft-100564050-orig.png

Is your graphics card overclocked (or factory overclocked) and have you tried lowering the clocks a little? The current clocks might be a bit too much for it, either from higher-than-normal heat buildup, or perhaps aging of the components. You might also want to check the card's temperatures, and verify that the card is not covered in dust, which could impede cooling performance, and that its fans are working correctly.

Edit: And I see someone already answered similarly. That's what I get for opening multiple threads at once and not reloading them before posting. : D

 

[QwerkyPengwen]

The only time I've ever seen this is when overlcocking the memory on a GPU to an unstable level. At first you get no artifacting and if the load on the memory isn't very high every time you use the GPU then it can take a while before it all of sudden shows up. But it's directly related to an unstable overclock of the memory. At least, as far as I've seen in all my years with computers and overclocking.

If you're not overclocking then it's still most likely the memory and it's going bad. And if it's the case then underclocking your memory can help reduce this issue if not solve it, but this will only be a bandaid not a cure and at any given moment your GPU memory could fail completely and render your GPU useless.

So to wrap up and summarize:

If you're overclocking memory, reduce overclocks until problem disappears.

If not overclocking, then underclock until it goes away and seriously consider getting a replacement or upgrade.

If not overclocking and then underclocking doesn't work, possible that your memory is still effed, but could also possibly be something else (although I doubt it) and you should come back here with a report either way.

 

[JWildOnes]

Sorry for the late reply. I have not touched the GPU clock at all. Everything is set to base clock speeds. I did set my ram speed to a higher rate in the bio to match the sticks capabilities, would that cause this?

I did a stress test yesterday for 5 minutes and reached a temp of 79 on the gpu.

I will try a new cable and monitor and see if that fixes things. But this may take a while as this only ever happens like once every 6 hours or so.

 

[audiospecaccts]

open up the computer, remove all the cards, then take a screwdriver and loosen and tighten the mounting screws going across. Sometimes if your motherboard mounting screws are not tighten evenly, the motherboard will warp.

Also check to make sure that all the screws are installed because some of the screw holes are the ground plane shielding and need to be installed.

 

[panathas]

Sorry for the late reply. I have not touched the GPU clock at all. Everything is set to base clock speeds. I did set my ram speed to a higher rate in the bio to match the sticks capabilities, would that cause this?

I did a stress test yesterday for 5 minutes and reached a temp of 79 on the gpu.

I will try a new cable and monitor and see if that fixes things. But this may take a while as this only ever happens like once every 6 hours or so.

No your PC system's main RAM cannot cause this. On the other hand the VRAM which is the graphics card's memory can certainly do, if it's starting to fail. That's why you should try to under-clock it (lower its frequency). You should try to do the same with the GPU (graphics processor) frequency, if the VRAM under-clocking has no effect. The most probable scenario is that your graphics card is failing though and the only thing you can do, is to prolong its life by lowering its frequency. However you should first test another monitor and cable, just in case. Good luck.

EDIT: Keep in mind that by under-clocking the GPU/VRAM, the GPU will lose some of its performance and the more you keep lowering the various frequencies the slower your graphics card will become.

 

[cryoburner]

Yeah, I wouldn't expect a big performance hit from underclocking the VRAM though, unless you had to significantly lower it. Most cards come somewhat overclocked from the factory compared to the reference design, and it generally only amounts to a few percent more performance or so. If you could run your 970 at reference memory clocks, I doubt you would be able to perceive the performance difference.

 

[audiospecaccts]

I would test to see if it is the video card since the video ram could do this too. I would just substitute with a cheap video card to see if it shows up.