Game crashes only with high GPU Usage

[Tammo1]

Hi there,

 

im here today, because i have a Problem with my System. I really hope that you guys can help me fix it. Im not the worst at assembling/fixing computers, i built and upgraded my rig since 2016 & built many different Computers for Friends & Family. I think i know which components are bad in my case, but maybe im wrong?

 

My Problem is, that i cant play any games which will max out my GPU. If my GPU hits the like 90+ % usage for a few minutes (60 seconds and up) my game just crashes with a black screen. In most games with no error code. But in Titanfall 2 i get this error code: "dxgi_error_device_hung". At the moment its not bothering me that much, because i only Play Cs2 with friends, where my CPU is bottlenecking my GPU (40-70%max). But sometimes i like to just play other games which requiere more GPU performance. If i limit my fps, so my gpu doesnt get high usage, everything is working fine!

 

Here are some things i already tried:

- DDU my Drivers

- Use older driver versions

- Use Windows 10 & 11 (please no linux :))

- Installed windows for scratch up

- refitting & dust cleaning my GPU & PSU

- Used the other Power PCIE Cables from my PSU

- i checked temps: GPU max 80C´ % CPU max 57C´ under full load.

 

And here is my Spec list:

- Ryzen 7 2700x

- 16GB DDR4 3200mhz (Crucial Ballistix)

- Gigabyte Aorus B450 m 

- Fractal Design Celsius+ S24 Prisma

- inno3d GTX 1080 i chill x3

- be quiet! System Power 9 600w

- 1x Samsung evo plus m.2 SSD, 2x Kingston 480Gigabyte Sata SSD, 1x some Random Western Digital 1tb HDD.

 

To conclusioni came: Either my GPU or PSU is faulty. I wanna try putting my GPU in a Friends PC to check if it has the same behavior there too. But for the Moment im restricted to this. Keep in Mind that the system was running fine for years in this configuration and the Problem appeared just a few Months ago. What i think is strange, that if i reinstall windows it will just work for like a week, even using my GPU at max. After that is just gets the same behavior.

 

I thougt about upgrading my GPU or PSU. I dont really need an upgrade, because under normal circumstances my rig has a perfectly good perfomance for my Usecase.

 

I am sorry for my Bad English, i am not a native English speaker, but i tried my best.

 

Thanks for your Help!!!!

[Sawa Takahashi]

Well, it is possible that the thermal paste has gone dry on your GPU and is causing thermal shutdowns. In that case, you should high temperatures on your GPU but since you see 80C max it should still be good to go.

The PSU going bad is more likely in that scenario. Even if it still works well most of the time, maybe its capacity has fallen just below what your computer needs at max GPU usage, or the voltage drops just a bit too much at high loads.

Most of the stores will allow to purchase a PSU and return it for a full refund (ask before just to be sure), so that is the easiest component to swap for a test.

Hope that helps.

[Tammo1]

5 minutes ago, Sawa Takahashi said:

Well, it is possible that the thermal paste has gone dry on your GPU and is causing thermal shutdowns. In that case, you should high temperatures on your GPU but since you see 80C max it should still be good to go.

The PSU going bad is more likely in that scenario. Even if it still works well most of the time, maybe its capacity has fallen just below what your computer needs at max GPU usage, or the voltage drops just a bit too much at high loads.

Most of the stores will allow to purchase a PSU and return it for a full refund (ask before just to be sure), so that is the easiest component to swap for a test.

Hope that helps.

Thanks for the quick answer. I renewed my GPU Thermal paste in summer 2022, bc it was running very hot. I think I’ll try the refund “method”. Good thing is: In my country store are obligated to take back purchased items up to 14 days with no questions asked. 

[Tammo1]

23 minutes ago, Sawa Takahashi said:

Well, it is possible that the thermal paste has gone dry on your GPU and is causing thermal shutdowns. In that case, you should high temperatures on your GPU but since you see 80C max it should still be good to go.

The PSU going bad is more likely in that scenario. Even if it still works well most of the time, maybe its capacity has fallen just below what your computer needs at max GPU usage, or the voltage drops just a bit too much at high loads.

Most of the stores will allow to purchase a PSU and return it for a full refund (ask before just to be sure), so that is the easiest component to swap for a test.

Hope that helps.

Do you got any recommendations for a PSU? I don’t wanna spend 120+€ but still don’t want some PSU that will explode (said exaggeratedly)

[Sawa Takahashi]

On 10/30/2024 at 2:27 PM, Tammo1 said:

Do you got any recommendations for a PSU? I don’t wanna spend 120+€ but still don’t want some PSU that will explode (said exaggeratedly)

Prices certainly won't be the same (I'm in Canada) but I got good experience with Seasonic, EVGA, Corsair or even the cheaper Thermaltake.

[Blue4130]

To me this sounds more like a dying GPU than PSU. My bet is on the VRAM going bad. Do you have another system that you can test the GPU with? Any friends nearby that you can borrow a system from?

[tunassub]

On 10/30/2024 at 2:27 PM, Tammo1 said:

Do you got any recommendations for a PSU? I don’t wanna spend 120+€ but still don’t want some PSU that will explode (said exaggeratedly)

Corsair makes great products in my experience and the RM650 is €84.90. 

[BetteBalterZen]

8 hours ago, Blue4130 said:

To me this sounds more like a dying GPU than PSU. My bet is on the VRAM going bad. Do you have another system that you can test the GPU with? Any friends nearby that you can borrow a system from?

A dying GPU usually shown signs like graphics glitches. OP does not mention any graphics glitches or errors. 

Personally, I suspect the PSU. Seems like when the GPU pulls the most power, crashes happen. 

[Blue4130]

35 minutes ago, BetteBalterZen said:

A dying GPU usually shown signs like graphics glitches. OP does not mention any graphics glitches or errors. 

Personally, I suspect the PSU. Seems like when the GPU pulls the most power, crashes happen. 

Not always. I have a dying 2080 in my system right now. No glitches, but as soon as the vram gets loaded up, it black screens and the monitor looses signal. The rest of the PC happily works in the background. I can hear the audio of any thing I have running. Tested by using a vram testing program, and by swapping my GPU with my wifes GPU. Whatever machine I put my card into does the same thing.

 

@Tammo1 Use a program like memtestg80 and see what happens.

https://github.com/ihaque/memtestG80

[Mark Kaine]

yea corsair rm/x650 is a good choice... bequiet pure power could be alright too for example (cost 55 euros here) i had the pure power 10 500w, could even run my 3070 without issues...

 

 

while we don't know, that system power psu is sus as hell so you honestly should replace it anyway. 

 

 

[Blue4130]

44 minutes ago, Mark Kaine said:

while we don't know, that system power psu is sus as hell so you honestly should replace it anyway.

What makes it sus? From the limited reviews I could find, it was decent, nothing special, but not anything to worry about.

[Mark Kaine]

 

edit: wrong topic