GPU intensive games crash under load

[PayZuni]

OS: Windows 10 Pro 64-bit

Motherboard: MSI B450 Tomahawk Max  (BIOS: American Megatrends International, LLC. 3.A0, 03.02.2021)

CPU: AMD Ryzen 7 3700X (Running stock)
RAM: 2x 8GB Corsair Vengeance LPX 3600MHz (Tried with and without A-XMP)

GPU: Galax Geforce GTX 980 SOC

PSU: Corsair VS550
Storage:

• Samsung SATA SSD 860 EVO - 1TB
• NVME M.2 Crucial P1 - 500G
• Toshiba DT01ACA100 HDD - 1TB

Problem:

When playing Games, which put load on the GPU, these Games sometimes crash with a message like "D3D Device Lost". When looking into Event Viewer there are 2 Messages that stick out to me:

• Error - Source: nvlddmkm - ID: 13 - No description
• Warning - Source: Display - ID: 4101 - display driver nvlddmkm stopped responding and has successfully recovered

 

 The GPU mostly runs at 79°C. To me this looks similar to:

They came to the conclusion, that it's probably the PSU which isn't powerful enough, but I really wanted to get a second oppinion before going out and buying a new PSU as I don't have a second one lying around for testing.

The only way I was able to get it to run stable was to lower GPU core clock and GPU Memory Clock using MSI Afterburner.

Thank you.

[Worstcaster]

The error messages seem driver related.  You could try running DDU and reinstalling the drivers (It costs nothing to try while you are looking for new PSUs).  That being said the PSU does seem weak for that system.  Nvidia is recommending a 650 W PSU as a minimum for a 3070.

 

[Levent]

This also could be unstable components on your graphics card (particularly VRAM and VRM). Can you reproduce this issue when you run Furmak alongside prime95 instantly or within 10-20 seconds?

[PayZuni]

2 hours ago, Worstcaster said:

The error messages seem driver related.  You could try running DDU and reinstalling the drivers (It costs nothing to try while you are looking for new PSUs).  That being said the PSU does seem weak for that system.  Nvidia is recommending a 650 W PSU as a minimum for a 3070.

 

I will try reinstalling the drivers using DDU later, even though I don’t think this will help much because I already wiped my entire System once and the error persisted. Also I am not running a 3070. The user in the other post had a 3070. I am actually running a 980.

 

1 hour ago, Levent said:

This also could be unstable components on your graphics card (particularly VRAM and VRM). Can you reproduce this issue when you run Furmak alongside prime95 instantly or within 10-20 seconds?

I will try that out later today.

[PayZuni]

7 hours ago, Levent said:

This also could be unstable components on your graphics card (particularly VRAM and VRM). Can you reproduce this issue when you run Furmak alongside prime95 instantly or within 10-20 seconds?

Okay ran Furmark and prime95 for a few minutes. It seemed to handle those just fine (Even though it got pretty hot, about 83°C on the GPU), but no crashes and no new entries in the Event Viewer. So maybe it actually is a driver problem?

[Levent]

1 minute ago, PayZuni said:

Okay ran Furmark and prime95 for a few minutes. It seemed to handle those just fine (Even though it got pretty hot, about 83°C on the GPU), but no crashes and no new entries in the Event Viewer. So maybe it actually is a driver problem?

That test was to check if your PSU was the problem or not, if you were run those tests at the same time if your PSU was not capable of handling it, it would have crashed instantly. What I would try next is, to have a more aggressive fan curve while gaming to see if that helps (or simply max out your GPU fan and see if it helps). Drivers are a possibility, so is Microsoft update, as there was one problematic update recently.

[PayZuni]

2 minutes ago, Levent said:

That test was to check if your PSU was the problem or not, if you were run those tests at the same time if your PSU was not capable of handling it, it would have crashed instantly. What I would try next is, to have a more aggressive fan curve while gaming to see if that helps (or simply max out your GPU fan and see if it helps). Drivers are a possibility, so is Microsoft update, as there was one problematic update recently.

Well I've had that problem for about half a year now, already. I will try to adjust my fan curve now and come back to you if the issue persists.

[PayZuni]

11 hours ago, Levent said:

That test was to check if your PSU was the problem or not, if you were run those tests at the same time if your PSU was not capable of handling it, it would have crashed instantly. What I would try next is, to have a more aggressive fan curve while gaming to see if that helps (or simply max out your GPU fan and see if it helps). Drivers are a possibility, so is Microsoft update, as there was one problematic update recently.

Setting a very aggressive fan curve and maxing out my GPU fan doesn't seem to resolve the issue. 

[PayZuni]

On 5/3/2021 at 12:02 PM, Worstcaster said:

The error messages seem driver related.  You could try running DDU and reinstalling the drivers (It costs nothing to try while you are looking for new PSUs).  That being said the PSU does seem weak for that system.  Nvidia is recommending a 650 W PSU as a minimum for a 3070.

 

Tried this today. Sadly it didn't resolve my problem.

[PayZuni]

Today my PC crashed with a BSOD. The error was "Page fault in nonpaged area". Maybe this is another indicator to where my problem lies? I really need help. My PC has become pretty much unusable at this point and I have no idea what to do.