Game causing computer to hard reboot

[event_horizon]

PC specs:

• CPU: AMD Ryzen 7700X
• GPU: Sapphire Radeon RX 5700 XT (Sapphire recommends a 650W+ PSU)
• Motherboard: Asus ROG STRIX B650E-I
• RAM: 32GB G.SKILL Flare X5 DDR5-6000
• PSU: 700W 80 Plus Gold (unknown brand, came with Hyte Revolt 3 case)
• CPU Cooler: Cooler Master MasterLiquid ML240L V2 RGB (240mm AIO)

 

I just built the computer in March with brand new components save for the GPU which I bought used of eBay. The GPU was in excellent condition when it arrived. I am using the Steam version. Pastebin of the last time the the computer rebooted.

 

When I first installed the game, I was able to run it at near max settings with decent frame rates. The game run for about five minutes and then my computer hard reboot. No blue screen, it would just shut down and restart. I lowered all of the settings to the minimum as then able to get about 10 minutes of playtime. I set the CPU to 65W eco mode and limited the GPU to 50% power in the AMD Adrenalin software and am now able to get 30~45 minutes of playtime (though sometimes less).

 

Windows event viewer show a Kernal Power 41 error.

 

Troubleshooting steps I have taken:

• Verified integrity of game files in Steam
• Removing and reinstalling video drivers via AMD's driver removal tool
• Making sure BIOS is up to date
• Running CHKDSK, SFC and DISM (I used this guide)
• Stress testing the PSU via OCCT - no issues
• Running MemTest86 - no issues
• Running FurMark for over an hour - no issues

 

The last time the computer rebooted, it corrupted my save file. I was on day 30. It's extremely frustrating to lose all that progress. I do not have issues in any other games.

Edited July 4, 2023 by event_horizon
Added troubleshooting step (see edit history)

[Beelzybub]

On 7/3/2023 at 8:52 PM, event_horizon said:

No blue screen, it would just shut down and restart.

This is a symptom of a failing PSU. An overheating cpu might do it also.

Try running prime 95 and Furmark at the same time and see what happens.

[event_horizon]

23 hours ago, Beelzybub said:

This is a symptom of a failing PSU. An overheating cpu might do it also.

Try running prime 95 and Furmark at the same time and see what happens.

I ran both for an hour, and my computer didn't reboot. The CPU and GPU were both hitting ~95°C.

[Jugginator]

Your Event Viewer logs will show more of what happened, but two things: Both your cpu/gpu hitting 95c is a sign that there's a severe lack of cooling, you shouldn't be hitting temps that high, and with that system, an unknown brand of PSU is sketch and I would gander that may be the issue. Generally, I wouldn't use a PSU that came with a case, unless the PSU was a common/major brand. I did that exactly once and it fried my motherboard and everything connected to it in less than a year - not saying yours will do that, but those included PSU's are the bare minimal quality (the gold rating is not an indicator of quality)

 

Oh, another thing I noticed, you're likely at the max draw of the power supply, that system alone not including any fans/storage devices/rgb or anything is about 700 on heavy load, which generally you don't want to go near the max the PSU can do, nevermind a no-named included in the case PSU.

Edited July 6, 2023 by Jugginator (see edit history)

[unholyjoe]

the scariest thing i noticed is this

"CPU Cooler: Cooler Master MasterLiquid ML240L V2 RGB (240mm AIO) "

 

any AIO that gets that hot is not working as it should. not all AIO are good and or even work, but at the least the temps should have been below 50c.

 

so yeah i am with jugg when he points out there is a cooling issue..

[event_horizon]

I pulled the PSU out of my old gaming computer (XFX 750W 80 Plus Gold) and it fixed the problem. I figured a brand new PSU shouldn't be causing the issue, but I was wrong.

 

For reference, the Ryzen 7000 CPUs are designed to run hot, so 95c is actually normal.

 

Thank you for the suggestions and assistance.

 

 

[Jugginator]

1 minute ago, event_horizon said:

I pulled the PSU out of my old gaming computer (XFX 750W 80 Plus Gold) and it fixed the problem. I figured a brand new PSU shouldn't be causing the issue, but I was wrong.

 

For reference, the Ryzen 7000 CPUs are designed to run hot, so 95c is actually normal.

 

Thank you for the suggestions and assistance.

 

 

 

Glad you solved it, and a reason why I always suggest people keep old/replaced parts that are known good for cases like this. I would recycle that PSU and never put it in a system lol. The cheap PSU's like that aren't built for quality, and hope that whoever uses them doesn't use anything but 50 - 60 percent of the total power capacity, since higher usage will be unstable. High quality PSU's can handle being near or at max load, lesser ones cannot.

[event_horizon]

Well, I bought a new PSU (Corsair SF750) and the problems persists. If stress testing the hardware doesn't cause the computer to reboot, it has to be a problem with the game.

[meganoth]

But you said your old PSU fixed the problem, so somehow it has something to do with the PSU, doesn't it?

 

Different PSUs often have their wattage distributed differently over the voltages. So one might have a beefier 5V line and less on the 12V line or vice versa. Did you check if maybe the XFX PSU was particularily strong in any one of the lines where the other two were not?

 

[Jugginator]

Try turning off SS Reflections and maybe SSAO, we do know of some issues related to certain hardware/driver configurations, but typically, blue screens/reboots happen when the system/hardware were asked to do something they couldn't handle at the moment. Might be worth getting BlueScreen from the sysinternals tool set and have a look or post here what the information is in it.

[Fox]

11 hours ago, event_horizon said:

If stress testing the hardware doesn't cause the computer to reboot, it has to be a problem with the game.

Wrong! Stress tests usually only stress 1 or 2 pieces of hardware at a time, not the entire computer as a whole. 7d2d stresses everything all at once. Don't be so quick to blame a game that has been around for 10+ years, especially when you're the only one having this issue right now.

 

GPU should never reach 95 degrees... ever. CPU reaching that temp, while claimed to be normal according to AMD, is still stupid and not ideal for reliability reasons which is why there's so much hate for that LGA CPU right now. Also, 95C is pretty much the thermal throttle limit, so while it may seem normal (again, according to AMD), you don't even know if it's performing like it should.

 

In any case, what computer tower are you using, because it's pretty obvious you're not getting enough airflow for that GPU (and probably CPU as well). With that kind of heat, who knows what damage and instabilities you could be dealing with. Keep in mind that ram, ssd, motherboard chipsets, motherboard power distributers, etc all don't have sensors to let you know they're hot. So if all the things that do have sensors are telling you everything is hot... then odds are, other things inside the case are really warm or hot too.

Edited July 19, 2023 by Fox (see edit history)

[FinkPloyd]

I would be very concerned if my GPU was running at 95C.

 

Isn't the Hyte Revolt 3 a mini ITX case?

If so I would be also watching the temp of your RAM (with someting like CPUID HW monitor).

People report that GSkill DDR5 has some heat management issues which are exacerbated with small form factor cases (due to the lack of air flow).

 

 

[Jugginator]

Yeah that GPU should not be that hot. And I will reiterate that stress testing isn't always indicative of stability. People like gamers nexus do wonderful work but they don't push the fact that stress testing and benchmarks are not exactly good indicators of stability for general usage, like playing a complicated game. I was too pushy with my undervolting/overclocking with my CPU when I first got it (used the new automatic curve optimizer from AMD and set my core clocks higher on boost) and it would sometimes crash with 7 days and another game, think it was csgo. Previously, I ran all sorts of stress testing for hours straight and no issues. Load up into 7 days or csgo and it may crash after 10 - 30 minutes or after a couple of hours. I bumped up the voltage to my 2 main cores by a tiny bit and it's been stable ever since. So, in summary, stress tests don't always hit the breaking point.

 

 

[SylenThunder]

As I have stated previously in another thread regarding PSU failure..  I could run Prime95 and Furmark side-by-side for days and not stress the system near as much as just running the game client. It did end up being the PSU in the end. Bad rail wasn't suppling enough power to the disk when there was a certain amount of load. Finally figured it out by running OOCT for a couple of hours and testing the actual voltage on the rails.

 

My bad using was a Corsair 750. Ended up replacing it under warranty with another Corsair, which ended up failing in a similar fashion, except it took out the SSD and almost fried the motherboard in the process. Replaced that with an MSI A850GF, and haven't had any issues yet.

 

All information provided so far points to either PSU failure, or thermal shutdown.