PC crashes while gaming - is this GPU issue? (Solved)

[happycats]

My PC crashes whenever I start gaming. Played Cyberpunk and Xcom2, and after a while, PC will randomly restart, usually within 5-10 minutes

I tried to recreate the issue by running the Heaven benchmark, and it also does the same thing, after a while, PC will reboot, also usually within 5-10 minutes

 

The following is what I've done to try and fix the issue:

1) Ran Prime95 CPU test --> PC runs fine even after 30 minutes of CPU torture test

2) Ran memtest86 --> Passes all the memory tests

3) Ran Heaven benchmark --> PC crashes and reboots within 5-10 minutes

4) Ran "WhoCrashed" --> I get the error: WHEA_UNCORRECTABLE_ERROR in the dump analysis. Some kind of hardware error it seems, but doesn't say which hardware.

5) Downloaded MSI Afterburner and reduced the Powerlimit to -50, and PC becomes stable, I can game on it without issues.

 

I also noticed that when I took my side and top panel off the case, the crash takes longer to happen, around 30 minutes rather than 5-10 minutes, which makes me suspect that it might be temperature related.

I checked to ensure that all the GPU, CPU and case fans are all running so it's not a problem with the fans not spinning.

 

So then I downloaded hardware info to monitor the temperatures of CPU and GPU while doing the stress tests above.

 

Running Heaven benchmark with case panels on, at the point of crash:

    CPU (Tctl/Tdie) [°C]:72.4

    CPU Die (average) [°C]: 68

    CPU CCD1 (Tdie) [°C]: 70.8

    GPU Temperature [°C]: 74

    GPU Memory Junction Temperature [°C]: 86

    GPU Hot Spot Temperature [°C]: 92

 

Running Prime95 CPU stress test with case panels on, at the point of crash:

    CPU (Tctl/Tdie) [°C]:77.7

    CPU Die (average) [°C]: 77

    CPU CCD1 (Tdie) [°C]: 75.8

    GPU Temperature [°C]: 51

    GPU Memory Junction Temperature [°C]: 52

    GPU Hot Spot Temperature [°C]: 51

    

Also when I lowered the Power limit via MSI Afterburner, and while gaming on it, I observe that the GPU temperature never got above 70 deg C, usually sat around 65 deg C. Even the Junction temperature measured by GPU Temperature 2, never got above 70 Deg. Power usage is just sitting around 90+ watts.

 

The following are my PC specs:

CPU: Ryzen 5600X with stock cooler

GPU: Gigabyte 5700XT Gaming OC

Motherboard: Gigabyte B550 Aorus Master 

Memory: 32 GB Corsair 

Power supply: Corsair RM750X 80+ gold 750Watt

 

I've updated the BIOS for the motherboard to: F11n

I've updated the GPU from the AMD website: Adrenalin 2020 20.12.1

 

At the moment, I'm suspecting I've got a dud GPU, although I read online that the GPU junction temperature, even if it's at 90 deg C, its considered acceptable? I'm not quite sure why it would crash at this temperature, even though I feel like it's temperature related because when I got my case panels on, it crashes a lot sooner. Is there any other temperature I need to be monitoring? Any other thing I'm missing out in my analysis ?

 

Thanks for your help

 

 

 

 

 

 

[Fragment323]

2 hours ago, happycats said:

Running Heaven benchmark with case panels on, at the point of crash:

 

2 hours ago, happycats said:

 GPU Memory Junction Temperature [°C]: 86

    GPU Hot Spot Temperature [°C]: 92

You found the problem: GPU temperature too high. Rest of the system look fine from what you showed here.

Problem can be a) cooler, b)defective GPU, c)heavily overclocked GPU

Now what you can do:

- Provide better cooling for GPU - case fans for intake and exhaust for example

- Disable any overclocking on the GPU

- RMA the GPU - if under warranty

- If out of warranty: check if GPU cooler makes good contact with GPU (disassemble and assemble it again, check if cooler makes contact with GPU and other components, put in new thermal compound)

- Get aftermarket cooler or watercool your GPU

- Get new GPU

[happycats]

Thanks, @Fragment323 I wasn't sure whether those GPU temperatures were too high or not, as I read in some other sites that it's supposedly normal.

Thanks for your feedback

[BiL]

Do you have a spare GPU to test or can you try yours on another rig?

[happycats]

@BiL unfortunately I don't have another GPU to test, however I brought it to a computer shop for them to diagnose the problem as I'm still not convinced it's the GPU, but was unable to reproduce the problem at the shop, but when I brought it home, it starts crashing again. Very strange, but I will keep testing and trying, and if I manage to get to the bottom of it, will post an update

 

[BiL]

On 12/26/2020 at 6:03 AM, happycats said:

@BiL unfortunately I don't have another GPU to test, however I brought it to a computer shop for them to diagnose the problem as I'm still not convinced it's the GPU, but was unable to reproduce the problem at the shop, but when I brought it home, it starts crashing again. Very strange, but I will keep testing and trying, and if I manage to get to the bottom of it, will post an update

 

Since the problem wasn't reproduced at the workshop we left with 3-4 more things to check and that's your keyboard, mouse, monitor and hdmi/DP cable that you're using (assuming that you went to the workshop only with your main unit). Finally, this may sound silly to some, but try to change position to your PC (another room and specially another wall plug)

[happycats]

I have finally got the solution to the problem.

I took it to another computer shop and they swapped out the GPU and RAM and found that it wasn't the GPU or RAM causing the problem.

 

The issue was the stock Wraith Stealth cooler that came with my AMD Ryzen 5600 CPU. 

It looks like the cooler is not up to the task, 

I have since upgraded to an AIO water cooler, and it has performed perfectly with no more crashes.

 

Strange thing though is that the CPU temperature monitoring that I used in Hardware info shows that the CPU temps only reaching around 70 deg, but looks like that's enough to crash the system.

 

Thanks for your help and responses

@BiL