PC Shuts down only with Unreal - TwinMotion

Hi!

I’ve built a PC for architectural visualization.

CPU: Ryzen 9 3900X

GPU: MSI Trio 3080 (new, intalled)

MoBo: ASUS Pro WS X570-ACE

PSU: Focus+ SSR-850FX

Built it, tested it (cinebench, 3d mark, multi hour gaming, Flight Simulator) all good, all nice.

One day, wanted to fill a landscape with trees in TwinMotion, 10 trees and it turned off suddenly.

Ok, tried again the trees, same thing.

Then, tried Unreal 4.27 on this PC, boom again shuts itself down. Sometimes while dragging a reflection sphere, sometimes when changing a material, some times rendering light on preview, or in production. It is jus dissapointing. Kernel Power event id 41.

Cant figure out what is UE doing to shut down my pc.

Any ideas?

I even lost a GPU it was an EVGA 2080Ti (burnt, total loss) magenta lines all over display no memory detect, etc).

Coukd it be overheating?

I believe it’s overheating issue.

I recommend cleaning your P.C.
Check your cooling system and turn off you’re overclocking.

Hi Limanima, thanks fot the reply.

My temps on CPU always below 80 C on a Artic AIO (temps checked via Ryzen Master and HwMonitor).

My case is a Fractal Meshify 2 and even bought 2 more 140mm fans, totalling 3 140mm front intakes and 4 exhaust). Full fans kick in at 70C for the CPU. In my opinion temps are sweet!

Weird thing is that i have tried:

  • CineBench
  • Heaven
    -Aida 64
    -Halo
    -Apex Legends
    -Microsoft Flight Simulator

All of the above perfect, not a single issue, even at ultra high settings. But the issue appears on Unreal-Based apps (Twinmotion-Unreal Editor).

Hi Striker.x, thanks for the reply.

As mentioned above, my temps are below 80C for the CPU and GPU, Aida 64 30minute test, all good.

As it is a new PC the thing is shinny new! no dust, no nothing.

As soon as the problem went crticall I turned everything to default, MoBo settings, RAM Speed, GPU an CPU to default as well. Same thing.

Forgot to mention i have 32GB RAM Ripjaws.

Ive tried renderings without the glass panel, changed powercord, direct to wall outlet, etc.

But the issue continues only with UE Based apps.

Sounds like you have a puzzle to figure, but I do not think it is that tough a puzzle at that. Are you sure your power supply has sufficient wattage to power your whole PC even with your graphics card fully throttled up? I say this because you need to power up many parts and you can not afford to not be too low on power. GPUs do run hot, personally I aim for 80c max.

UE4 will try to run any project at the highest frame rate possible in the editor, this is why any project or scene will send the temps up and max GPU usage. You can try limiting the frame rate of the project you are working on or setting real-time to off in the viewport

You can also try under volting your GPU to substantially lower the power draw with only dropping the performance by a bit. To be on the safe side of things, check to see if you have enough power to support your system and double check to see if your GPU is not over heating which may be caused by over clocking (which you claim is not) or a power supply that is not working correctly.

Hi ANFETOMA,

Since your computer is in excellent condition.
But I still recommend monitoring your power supply, ram, CPU, GPU, in idle mode…
And Fully turn off your antivirus while editing.


Maybe it is a Twinmotion bug(s).
Even UE5, Twinmotion bug(s) still exist.
Unfortunately, I cant help you in this matter.

Try running OCCT for 10m on both CPU and RAM.

Sudden power loss without a bsod indicates a hardware issue.

Could also be driven by something like a liquid cooler water temperature exceeding a certain threshold.

Hi, thanks for the reply.

I have a MSI TRIO 3080, which recommends 750W PSU and consumes 340W. Now I have a Seasonic Focus+ 850W PSU, I think it should be sufficent. Im no expert.

I did try to lower FPS to 60fps and also turned off the Real time Viewport, same thing.

But the thing is that it happens all over Unreal, I mean, shuts down on material editing, rendering, dragging lights to the scene, connecting nodes. And the crazy thing is that ive tried to recreate the event and it does not happen.

For expample, one day tried rendering on production lightinig it shut down. Restarted the PC, opened the project hit render again, completed 20min render without problem. Sometimes lets me work hours and sometimes happens few minutes after starting the editor.

Is there anyway I could record the data of my PC and after a shut down try to figure what trigger it? Windows Event Viewer says nothing rather that windows sut down unexpectedly.

Have you tried geforce studio drivers over the geforce one?

I’ve run 3D Mark, Aida 64, Heaven and now OCCT. They all work perfectly, no problems.

It only happens with Unreal or TwinMotion. As mentioned, other games or software no problems.

How can I know of the water threshold is been exceded?

Thank you for your reply.

I have monitored all of the components, nothin abnormal. The only confirmed issue is that only happens on Unreal software (editor and twinmotion).

Ive tried turning off McAffe but to be honest, i dont know if it turns off 100%. I turned off Firewall and real time scan.

With my past card and this new one, Ive had the latest studio drivers. Same problem.

Unreal likes to max out FPS, so you should limit it (to test if that is problem). When unreal goes for max FPS it pushes GPU to the limits, then it throttles (power consumption or overheating). I have similar setup (2080TI) and i had some random reboots.

If that does not work, i would get at least 1000W PSU, and quality one.

Other than that:
motherboard can have damaged power lines (when your 2080TI got burned it could damage something on motherboard)
get UPS it all may be caused by neighbors old washing machine.

Nawrot, Thanks for your time and reply.

Im considering a new PSU, but is it normal to need a 1000W PSU for Unreal Engine?

Im considering a new MoBo too hahahaha.

Ive had a problem that forgot to mention. Since day one, randomly my MoBo showed the “wait for RTL8117 kernel ready”. And even when i turned off Windos via “Start-shut down” it would not turn on again unless I flipped the switch at the back of the PSU. But lately the message went away and PC booted normally.

Any suggestions on specs for a UPS?

Capacitors in PSU will lose capacity over time, so you should assume that PSU loses about 5-10% of power per year. It is nice to have PSU with some extra power, it will also be quiet.

Not sure about UPS there are so many of them.

Thought about your problems:
If you can get whole PC to some service/repair shop and ask them to determine what is cause of problems, you may save some bucks by replacing only malfunctioning part, and not all of them.

PS.
I had similar problems in past, so since then i always have 2 PCs (current one, and previous/old).
I do not sell or give old parts, i keep them to find what is cause of such crashes etc. It also helps
to have spare pc (with lan drivers that work) in case system drive dies. Yes now i can boot linux from usb, but it is so easy to lose that usb.

To be fair - no, nothing a program will do will cause your system to just shut down unless that program is directly accessing something to do it, on purpose.

So, no.
Unreal will not “just shut down” your system.

A failure of the system is causing you to shut down.

Without a bsod?
Possibly a Bad gfx.

You mentioned you were able to run several test programs.
Assuming you tested both cpu and ram, and you are satisfied you get no errors after a 1 hour test.
Then start testing the graphics card.

First run GpuZ and set it to log things - because if you crash you may either bsod or need to hard reboot. Knowing where the files are stored is helpful.

Second, run Furmark for an hour on the best possible settings with the intent to cause the gfx to start a fire.

Usually, termps should remain Below 90c.

Because of how you described your problem, I don’t think the system will run fine for an hour.

If it does, then there’s something else that the matter.

And the only think you have left to test is your data storage…

As far as your other question. Some programs like iQue (corsair) monitor the water temperature and allow/cause a shut down to prevent damage.
I doubt there’s any such monitoring going on for a closed loop GFX, and the fact you ask likely means you aren’t running a custom loop, so it’s highly unprobable that’s an issue in your case.

Yes and no. I have 800W “gold” PSU here and it is kind of too low power for my hardware.
I have 2080TI, few SSD drives and 2 mechanical ones, 800W is bit lower than minimal recommended power.

Usually everything is fine, until is set any unreal based game to no VSYNC, then FPS goes up to 90-100 and sudden restarts happen. This same stuff happened since UT3 years ago, so by now i know to limit FPS for unreal based games, this does not let them go for max FPS and max power consumption.

… furmark would crash / shut you down too…
Also, why not just get $120 worth of a proper psu?

Thank you MostHost for the reply.

Yesterday I tried OCCT for an hour on CPU, it was supposed to be an extreme test…I dont know if Im making a mistake on OCCT test settings (did it on default settings set to one hour), but temps never went over 70C, nor the fans needed to kick in for the whole hour test…weird but im open tu suggestions on settings for the test to run it again. Planning to run Furmark tonight.

Ill keep you posted.