Freezing during gaming with artifacts.

[id104335409]

Lately I can't run any game on my PC without it freezing completely and causing me to manually reboot it.
What happens is I get visual artifacts during gaming like triangles accross the screen or giant words just flashing. These artifacts are surely parts of the text and subtitles of the games being thrown at the screen distorted. Other than that the games run fine without massive frame dropping or tearing or other objects or colors getting distorted. I tried on different games both high demanding and not so much. I thought it was strange that it is just text that randomly appears.
I havent seen any GPU or CPU spikes in temperature (always bellow 50C)or any lag. Besides gaming everything else just runs completely normal. RAM test shows OK. The drivers are upo to date.
Do you think it has something to do with the graphic's card? Like is it dying? It runs perfectly fine outside gaming. I watch BR rips, I surf the web.
Or perhaps a power deficiancy? As I said the GPU runs smooth through benchmarks.
DirectX12 issues? Drivers incompatibility? Power management options? BIOS settings?
My specs:
MB ASROCK Z87M Extreme 4
PSU Raidmax 735 Wats 80 Bronze
CPU Intel Core i7-4770 @ 3.40GHz watercooled
GPU Nvidia GeForce GTX 770 2GB RAM
RAM 8GB HyperX blu
SSD Samsung 840 EVO 120 GB
OS Windows 10 Pro x64
monitor is Asus 1080p @ 60Hz trough HDMI

I used to have random freezes when I first got this PC. That was solved with a firmware update on the SSD. Games were playing fine back then. This is something new. Windows updates maybe?

 

[Pat Flynn]

If you have another graphics card you can swap in to test, I'd give that a try. It sounds like the ram is starting to fail on graphics card. If you're still getting artifacts with a second card, I'm going to say motherboard- but start by trying another card to rule it out.

 

[DSzymborski]

Lately I can't run any game on my PC without it freezing completely and causing me to manually reboot it.
What happens is I get visual artifacts during gaming like triangles accross the screen or giant words just flashing. These artifacts are surely parts of the text and subtitles of the games being thrown at the screen distorted. Other than that the games run fine without massive frame dropping or tearing or other objects or colors getting distorted. I tried on different games both high demanding and not so much. I thought it was strange that it is just text that randomly appears.
I havent seen any GPU or CPU spikes in temperature (always bellow 50C)or any lag. Besides gaming everything else just runs completely normal. RAM test shows OK. The drivers are upo to date.
Do you think it has something to do with the graphic's card? Like is it dying? It runs perfectly fine outside gaming. I watch BR rips, I surf the web.
Or perhaps a power deficiancy? As I said the GPU runs smooth through benchmarks.
DirectX12 issues? Drivers incompatibility? Power management options? BIOS settings?
My specs:
MB ASROCK Z87M Extreme 4
PSU Raidmax 735 Wats 80 Bronze
CPU Intel Core i7-4770 @ 3.40GHz watercooled
GPU Nvidia GeForce GTX 770 2GB RAM
RAM 8GB HyperX blu
SSD Samsung 840 EVO 120 GB
OS Windows 10 Pro x64
monitor is Asus 1080p @ 60Hz trough HDMI

I used to have random freezes when I first got this PC. That was solved with a firmware update on the SSD. Games were playing fine back then. This is something new. Windows updates maybe?

Sounds like a failing GPU, to be honest; you start to see these things when the VRAM is going sometimes. Outside of gaming, you're not using much VRAM. I'd try gaming with the integrated graphics just to see if the issues repeat.

If you do end up needing to replace your GPU, I strongly urge you to take the opportunity and also replace the low-quality PSU so that your new GPU is being better protected.

 

[SkyNetRising]

What happens is I get visual artifacts during gaming like triangles accross the screen or giant words just flashing.

Can you make a photo of those artifacts?

 

[id104335409]

DSzymborski's comment is what is more frequent as an answer to these type of questions around the net. It is more likely the GPU's fault, unless someone has a direct solution.
I would change the GPU with a friend of mine's, but then I am pretty sure the PC would react differently to the games all-together... and then I am not sure what that would prove.
Artifacts look different in every game, but mostly triangles popping up outwards from the middle of the screen or broken mirror effect and giant words or letters in the left side of the screen. I am not sure how to record them...

 

[SkyNetRising]

I am not sure how to record them...

With your phone camera.

 

[id104335409]

I've been busy trying out different things
https://www.youtube.com/watch?v=5n3j9_OlPwg
I chose this one as it was record time breaking crash. Not much visible triangles tho. I had a lot of those in Mass Effect Andromeda, also shimmering, texture issues and light strobing. Rise of the Tomb Raider holds up to about 5 minutes with no glitches or artifacts and then crashes. But this one game can crash in less of a minute which is why I chose to make my tests on it.
So far I removed the graphics card and tried without it. The game runs, but barely - 1 to 2 fps. I tried a lots of BIOS changes some made the PC unusable, so I had to remove the CMOS a couple of times. Tried installing the game on the SSD to rule out HDD. I removed and installed the drivers completely fresh. Ran several MemTests to check the RAM. Ran several free benchmarks - they all pass within expected results. The GPU is not overclocked in any way. I was swapping power cables and HDMI cables. Still could be a PSU issue, but only shows in games...? And that damn game isn't all that demanding either. And the first level is just tiny.
All temps are always under 40 and the GPU is clean. All load and memory usage never ramps up.

 

[id104335409]

So I did a clean install on Windows 10 and unfortunately that had no effect either.

 

[SkyNetRising]

What's the manufacturer/model name of your card?

Try downclocking your GPU.
40C for GPU under full load? Doesn't sound right.

 

[id104335409]

It's a Palit
oh, no - the temps are not under 40 at full load. They do rise up to 70 under Furmark. It's just that when I follow the temps with monitors they are always low - the point is that the card is not overheating and it is not clogged or anything. The reason for freezing is something else. I am now investigating possible IRQ conflict.

EDIT I can't downclock the GPU or just don't know how. No such option in Afterburner, but I can undervolt it. In fact one of the many many time I tested different options I managed to keep the game running with no artifacts. But I couldn't replicate the process as hard as I tried, so it must have been a fluke.

 

[helpstar]

update your BIOS inside BIOS itself with EZ-Flash menu

try running 3dmark (it´s for free, the basic edition), after it click on compare result online and post the address which comes up in your browser

use intel processor diagnostics tool
use intel driver update utility

 

[Jai Molloy]

According to wikipedia:

"In some rare conditions, two devices could share the same IRQ as long as they were not used simultaneously. To solve this problem, the later PCI bus specification allows for IRQ sharing, with the additional support for Message Signaled Interrupts (MSI) in its later revisions. PCI Express does not have physical interrupt lines at all, and uses MSI exclusively."

So, as recommended before, try another card. The likelihood that it is an IRQ conflict is about as likely as the Earth being flat and made of lasagna.

I know this is not convenient but the order of troubleshooting no longer includes hunting down IRQ conflicts, I would troubleshoot in this order:

1. Driver Update (you probably have done this)
2. Windows reinstall (you already did this)
3. BIOS/UEFI (reset to defaults and try again. update, reset to defaults and try again)
4. Memtest (unlikely, but not a bad thing to try. Needs to be run for 24+ hours, I have had bad RAM test good in short runs)
5. Load test PSU with PSU tester (do not have it plugged in to anything but the load tester)
6. Different slot on the motherboard (contact in slot could be bad, not likely)
7. Different card (prime suspect, card is defective)
8. Different motherboard (the second shooter on the grassy knoll)

As you can see, I included a lot of steps, but you can save yourself a lot of time and go for the likely suspects (in this order), bad card, bad motherboard, bad PSU or bad RAM.

By the fact that you are running your monitor off of the card, it is not an IRQ issue. You wouldn't get that far into windows with that IRQ problem.

 

[DSzymborski]

Yeah, I don't think I've run into an IRQ conflict since Windows 95.

 

[id104335409]

Thanks Jay.
Yes, I have done all of these things. What I don't like about memtest is that it takes sooooo long to do something that a RAM swap can show in seconds. You swap it, you run a game - boom, you see if the damn RAM was the culprit.
I have this game that somehow shows right away if the GPU is having issues. I don't understand why Benchmarks and stress test can't do that. (the game is Battlechasers if anyone wants to know, AAA games like Tomb Raider can last for minutes with no visual signs)
I don't have anybody that close to be willing to swap RAM or GPU or PSU with me. Basically I have to do that in a PC shop. But that defeats the purpose of troubleshooting your own PC. Because I just have to give up and send the whole PC in town and pay someone to fix it. Worst case would be to have to buy a graphics card and trow away the old one. Best case - some obscure BIOS setting configuration and some smirks from the guys that work there. I would be fine if it is a PSU - everyone changes that from time to time.
The IRQ was something that I never thought about and that is why I made a second post, because I didn't even knew what am I watching at.

 

[DSzymborski]

And you've removed the GPU and tested things out on the integrated graphics?

 

[id104335409]

And you've removed the GPU and tested things out on the integrated graphics?

Yes, runs as expected - very low frames, huge lag. But it can't crash the PC.

 

[DSzymborski]

And you've removed the GPU and tested things out on the integrated graphics?

Yes, runs as expected - very low frames, huge lag. But it can't crash the PC.

Yup, I'm thinking you need a new GPU (and if you buy one, I'd get a better quality PSU to protect the new one). I believe if you've done most of these things already, you've exhausted how far you can go without replacing a part and taking your chances. I'm about 80% sure it's your GPU at this point (you can always get it from somewhere that will allow you to return the GPU).

 

[Jai Molloy]

Thanks Jay.
Yes, I have done all of these things. What I don't like about memtest is that it takes sooooo long to do something that a RAM swap can show in seconds. You swap it, you run a game - boom, you see if the damn RAM was the culprit.
I have this game that somehow shows right away if the GPU is having issues. I don't understand why Benchmarks and stress test can't do that. (the game is Battlechasers if anyone wants to know, AAA games like Tomb Raider can last for minutes with no visual signs)
I don't have anybody that close to be willing to swap RAM or GPU or PSU with me. Basically I have to do that in a PC shop. But that defeats the purpose of troubleshooting your own PC. Because I just have to give up and send the whole PC in town and pay someone to fix it. Worst case would be to have to buy a graphics card and trow away the old one. Best case - some obscure BIOS setting configuration and some smirks from the guys that work there. I would be fine if it is a PSU - everyone changes that from time to time.
The IRQ was something that I never thought about and that is why I made a second post, because I didn't even knew what am I watching at.

No problem man, when I first started out being a tech I had no spare parts, you just acquire them over time. Unfortunately I am virtually certain that the GPU is the culprit. Is there a flea market around your area? You don't need the best card, just a card even a lowly gt710 will prove the card is bad, but won't eliminate the PSU as a culprit because it doesn't draw enough power to stress the PSU.

 

[SkyNetRising]

EDIT I can't downclock the GPU or just don't know how. No such option in Afterburner

You change core clock increment to a negative value - that's downclocking.