PC Restarts While Playing Demanding Games

[AtomicVoyager]

GPU:  Gigabyte RX 7800 XT

CPU: AMD Ryzen 7 5800X

Motherboard: AsRock B550M Pro4

BIOS Version: P3.40 (1/18/2024)

RAM: 32GB DDR4 3200 with XMP enabled

PSU: Corsair 750e

Case: Corsair 4000D

Operating System & Version: Windows 10 (22H2 [19045. 4894])

GPU Drivers: AMD Adrenaline 24.8.1 (8/23/2024)

 

I just built this PC over the past few days. Recently, I have been playing less demanding games such as D2R. However, with the 1.0 release of Satisfactory, I wanted to jump back into that. I also like to play Tarkov. Satisfactory boots and runs just fine until I get out of the intro cutscene. After that, my screens go black and my computer restarts. I have tried multiple times and even got a BSOD once. Tarkov allows me to play a raid for about 20 minutes and then the same restart issue occurs. 

 

Any ARPGs or "less demanding" games seem to have zero issues other than some occasional frame time spikes. 

 

The CPU temp only gets to a max of 75C and the GPU at about 65C albeit at 100% utilization as the graphics are maxed (as set by the game's benchmark).

The PC outside of games runs just fine and boots quickly.

 

I have the most recent Windows updates and I have tried to update any and all drivers I can find whether that's the GPU, chipset, BIOS, etc.

 

Any advice would be incredible as I feel like I have hit a wall... Thank you!!

[TetraSky]

Sounds like you might need to stress test your hardware properly.

It could be anything from bad ram, bad motherboard, bad GPU. Or even bad windows install, bad driver, etc...

Usually, on a new build, the most likely culprit is the RAM.

 

I would strongly recommend running Memtest86+ (windows usb installer) for at least 24 hours (long time, I know. But you're having issues. If you were not, I'd say to run it for only 3 passes and call it good enough.)

 

After that, just run a stress test for the GPU, such as Furmark. If it crash and burn within 2 hours... It was bad. If it does not, great! Though this does not fully test "stability". Monitor your temperatures closely! 

https://geeks3d.com/furmark/

 

There's also OCCT that is an all in one stress/stability test tool. Highly recommended.

https://www.ocbase.com/occt/personal

Might be worth running a few times.

[downinthemorgue]

When your computer restarts, does it start back up on its own from the bios splash screen?

[Guest]

Sounds like PSU failure. Try a spare PSU or power limiting your hardware, and check all your connections, avoid things like daisy chain cables. That PSU should be a good unit (im assuming you mean RM750e), which shouldn't have problems but anything can fail/arrive defective. RMA if you can isolate it to the PSU.

It could also be memory, try turning off XMP. Higher capacities like 32GB can be heavier on the IMC depending on your config.

[AtomicVoyager]

23 hours ago, downinthemorgue said:

When your computer restarts, does it start back up on its own from the bios splash screen?

Yes

[AtomicVoyager]

On 9/16/2024 at 11:26 AM, TetraSky said:

Sounds like you might need to stress test your hardware properly.

It could be anything from bad ram, bad motherboard, bad GPU. Or even bad windows install, bad driver, etc...

Usually, on a new build, the most likely culprit is the RAM.

 

I would strongly recommend running Memtest86+ (windows usb installer) for at least 24 hours (long time, I know. But you're having issues. If you were not, I'd say to run it for only 3 passes and call it good enough.)

 

After that, just run a stress test for the GPU, such as Furmark. If it crash and burn within 2 hours... It was bad. If it does not, great! Though this does not fully test "stability". Monitor your temperatures closely! 

https://geeks3d.com/furmark/

 

There's also OCCT that is an all in one stress/stability test tool. Highly recommended.

https://www.ocbase.com/occt/personal

Might be worth running a few times.

I completed a 30ish minute stress test with Furmark and their CPU burner running at the same time. This allowed the CPU to reach 100% util at 90C. Nothing hitched or acted weird but I do understand the need to stress test for longer. I will start with the 24 hr RAM test as that seems to be the common response in this thread. I will update in about 24 hrs. Wish me luck.

[AtomicVoyager]

23 hours ago, rippy4500 said:

Sounds like PSU failure. Try a spare PSU or power limiting your hardware, and check all your connections, avoid things like daisy chain cables. That PSU should be a good unit (im assuming you mean RM750e), which shouldn't have problems but anything can fail/arrive defective. RMA if you can isolate it to the PSU.

It could also be memory, try turning off XMP. Higher capacities like 32GB can be heavier on the IMC depending on your config.

When I updated the BIOS, it removed the XMP setting and refreshed everything back to the defaults and yet I was still having the restarting issue. 

 

Yes, I have the RM750e, my bad.