GPU VRAM failure?

[SakuraBliss]

Hello. So a bit of backstory behind my issues. I stream to twitch using SLOBS. I mainly stream osu! (which has never given me any issues), but whenever I would try to stream a game such as Overwatch, my SLOBS would crash, along with all of the other programs running on my computer. After this happening to me many of times, I decided to look into what may be causing such issue. For the past couple of days I have been running test streams on SLOBS and messing around with certain settings. I would find myself having the same issue, osu! runs fine, but everything would crash when I would try to load up any kind of fps game. After messing around with settings for a couple of hours, I had thought that I had fixed my issue. I had uninstalled OBS, StreamElementsOBS.live, RazerCompanionApp, and NVIDIA Broadcaster (which I had been using for my mic). I disabled my steam wallpaper engine while running a stream and seemed to be able to stream these fps games fine now. Ensuring that my Windows GameMode was turned off with all of those other settings disabled, along with disabling any overlay that might have been running with the game seemed like the solution for me. That was until I got back onto my computer the next day to try and play COD Warzone. Trying to run this game caused some of the worst problems to my GPU. Attached below are some screenshots of some crash messages I have got, along with a video of what the crashes look like from my end (it was originally a Snapchat video I sent to my friend, so I apologize for the language). My drivers are up to date and I even uninstalled and reinstalled my GeForce Studio Driver. I haven't done any serious overclocking apart from using the Gigabyte EasyTune AutoTuning feature (which I am unsure even did anything). I have my specs posted below as well. Is this an issue that can be solved by myself, or is my GPU broken?

• GPU - EVGA GeForce RTX 2060
• CPU - AMD Ryzen 7 3700X 8-Core Processor
• Motherboard- Gigabyte GA-AB350-GAMING 3 ATX AM4 Motherboard
• RAM - 2x8GB Corsair Vengeance DDR4-2400
• Storage- 2x1TB WD Caviar Blue HDD
• Cooler- be quiet! Dark Rock 4 CPU Cooler
• Power Supply- Corsair RM750

 

 

 

 

 

 

Please let me know if you need any more information.

 

Thank you so much!

[CommanderAlex]

8 minutes ago, SakuraBliss said:

My drivers are up to date and I even uninstalled and reinstalled my GeForce Studio Driver.

Was this done by means of Control Panel/Settings or DDUing the driver? 

 

[AndreiArgeanu]

try stress testing your GPU with something like FurMark or MSI Kombustor, also try reinstalling your drivers by first DDUing and then installing the new driver.

[SakuraBliss]

1 minute ago, CommanderAlex said:

Was this done by means of Control Panel/Settings or DDUing the driver? 

 

This was done simply by reinstalling the NVIDIA GeForce Display Driver and selecting the factory reset option under advanced installation.

[CommanderAlex]

2 minutes ago, SakuraBliss said:

This was done simply by reinstalling the NVIDIA GeForce Display Driver and selecting the factory reset option under advanced installation.

I would download a copy of fresh NVIDIA drivers from online, download and install DDU, unplug the ethernet cable  to prevent Windows from automatically install a driver sourced from the internet. DDU the drivers, log back in and install the NVIDIA drivers you downloaded previously, then reconnect the ethernet cable and stress test the GPU using a benchmark program. 

[bmx6454]

try using ddu first in safe mode, then reinstalling current drivers. https://www.guru3d.com/files-details/display-driver-uninstaller-download.html

have you tried running the gpu underclocked/undervolted to see if it improves at all?

[SakuraBliss]

Just now, bmx6454 said:

try using ddu first in safe mode, then reinstalling current drivers. https://www.guru3d.com/files-details/display-driver-uninstaller-download.html

have you tried running the gpu underclocked/undervolted to see if it improves at all?

I have not tried running it underclocked, no. I will try using the ddu and reinstalling.

[SakuraBliss]

10 minutes ago, CommanderAlex said:

I would download a copy of fresh NVIDIA drivers from online, download and install DDU, unplug the ethernet cable  to prevent Windows from automatically install a driver sourced from the internet. DDU the drivers, log back in and install the NVIDIA drivers you downloaded previously, then reconnect the ethernet cable and stress test the GPU using a benchmark program. 

I have FurMark downloaded for a stress test, how long should I let the test run and what am I looking for? This is the first time I have done a stress test.

[bmx6454]

8 minutes ago, SakuraBliss said:

I have FurMark downloaded for a stress test, how long should I let the test run and what am I looking for? This is the first time I have done a stress test.

i would let it run for a couple hours at least. you are looking to see if it will still crash, or not, while trying to gather useful info to help diagnose the issue if it does crash.

[AndreiArgeanu]

8 minutes ago, SakuraBliss said:

I have FurMark downloaded for a stress test, how long should I let the test run and what am I looking for? This is the first time I have done a stress test.

You're looking for crashes basically or visual artifacts. So if the benchmark crashes, or your system restarts or shuts down as a result, or there are visual artifacts then that could mean that there is an issue with your graphics card.

[SakuraBliss]

7 minutes ago, bmx6454 said:

i would let it run for a couple hours at least. you are looking to see if it will still crash, or not, while trying to gather useful info to help diagnose the issue if it does crash.

 

7 minutes ago, AndreiArgeanu said:

You're looking for crashes basically or visual artifacts. So if the benchmark crashes, or your system restarts or shuts down as a result, or there are visual artifacts then that could mean that there is an issue with your graphics card.

ok. i just did the DDU uninstall and NVIDIA reinstall, I'll give the stress test a go now

[SakuraBliss]

19 minutes ago, AndreiArgeanu said:

You're looking for crashes basically or visual artifacts. So if the benchmark crashes, or your system restarts or shuts down as a result, or there are visual artifacts then that could mean that there is an issue with your graphics card.

 

19 minutes ago, bmx6454 said:

i would let it run for a couple hours at least. you are looking to see if it will still crash, or not, while trying to gather useful info to help diagnose the issue if it does crash.

Does this look normal so far?

[CommanderAlex]

2 minutes ago, SakuraBliss said:

~snipped image~

Does this look normal so far?

Looks good to me so far. 

[SakuraBliss]

Just now, CommanderAlex said:

Looks good to me so far. 

Are there any other tests you might recommend me running to ensure that my GPU is fine?

[CommanderAlex]

2 minutes ago, SakuraBliss said:

Are there any other tests you might recommend me running to ensure that my GPU is fine?

I sometimes do UNIGINE Heaven Benchmark, 3DMark like Time Spy/Fire Strike Extreme and also inside the Stress Tests that run for about 10 minutes that checks system stabilities. 

[SakuraBliss]

Just now, CommanderAlex said:

I sometimes do UNIGINE Heaven Benchmark, 3DMark like Time Spy/Fire Strike Extreme and also inside the Stress Tests that run for about 10 minutes that checks system stabilities. 

Let's say that, hypethetically, I run these tests and I don't see any instability, but I still have crashing issues like I originally posted... what would the issue be then?

[CommanderAlex]

4 minutes ago, SakuraBliss said:

Let's say that, hypethetically, I run these tests and I don't see any instability, but I still have crashing issues like I originally posted... what would the issue be then?

If its really suspected to be VRAM issues, then a RMA is in order for the graphics card if you still have time left on the warranty. Have there been any issues prior to beginning to stream? 

One thing you haven't mentioned that I'm not sure it would matter is how are your temperatures for CPU/GPU? 

[SakuraBliss]

6 minutes ago, CommanderAlex said:

If its really suspected to be VRAM issues, then a RMA is in order for the graphics card if you still have time left on the warranty. Have there been any issues prior to beginning to stream? 

One thing you haven't mentioned that I'm not sure it would matter is how are your temperatures for CPU/GPU? 

Temperatures are normal. And the issues seem more frequent after starting to stream.

[CommanderAlex]

3 minutes ago, SakuraBliss said:

Temperatures are normal. And the issues seem more frequent after starting to stream.

Have you checked VRAM memory usage before and after starting to stream? Does turning down the graphical settings make a difference? Use something like HWinfo64 to see if you are really maxxing out VRAM. 

[SakuraBliss]

The jpeg of the task manager I posted before shows what happened during a full SLOBS crash. Here is the image again-