Unity PC game causing a reboot?

I have a player of one of my Steam games with an issue I’ve not come across before.

He’s saying that when playing my game it randomly causes his PC to reboot, no popup, no warning, nothing. He sent me the log files and there’s nothing obvious in them. At first I thought it might be a hardware issue but he says he plays more taxing games fine and also ran some benchmarks without issue.

I really don’t know where to begin with this, it’s the first time I’ve encountered it in 12+ years game dev.

Any ideas guys? Thanks

  • Overheating on CPU or GPU. High heat can trigger emergency shutdown which can come with no warning at all.
  • Driver bug + faulty hardware. Could be any piece of hardware.
  • Malware.
  • Faulty power cable. Like the guy touches it with his leg while playing then boom.
  • Faulty PSU which can’t handle high CPU load.

Those are the most probably scenarios, in my opinion.

2 Likes

Thanks for the reply but why would any of those only happen when running my game?

This isn’t too uncommon, really.

Assuming the system is “only” 99% stable there’s always some application or a specific part of an application that triggers the 1% instability.

At the most extreme, this is why overclockers run Furmark and Prime95 for at least an hour. You could overclock a system to a level that seems very stable on regular use, but still crashes either of these two apps within minutes because they exercise the hardware in an extreme way.

Any application can cause such a hot loop. All it may take is repeatedly issuing the same tasks, or using a specific shader or CPU instruction or a specific sequence of these that overloads some internal transistors briefly, causing a failure that the hardware cannot recover from.

It’s like russian roulette. :wink:

2 Likes

If the game has vsync disabled it can easily bump up CPU or GPU use to the maximum. So that can trigger overheating, OR that can trigger PSU problem.

For example, on my PC I had PSU die when for fun I launched a software that created 100% CPU load on all my cores. It went like this: All fans immediately spin up and then boom. Dead PSU in 10 seconds. I guess I got my share of fun in the end replacing this thing.

Basically your user should run any tool that monitors temperatures and see if they go too high during gameplay. When temp starts going to 80 degrees Celsius on CPU or it is sitting at 75C while idling, that is likely to indicate that there is a problem. Thermal paste dried up, too much dust in radiator and so on. Shutoff can trigger at 95 degrees, but that’s motherboard specific. The user should also check if the problem occurs with any other intensive game. As far as I know, with GPU there’s no clear rules what happens in case of overheat. It could start artificing, it could just shutdown your PC, or it could bluescreen.

1 Like

Yeah, if the GPU is allowed to go all out bad things happen (especially if the card is an AMD one).

Right, and (hypothetically) a heavier-running AAA studio game with no vsync, still might have a big enough CPU load that will slow the rendering down to 60 frames a second and so the GPU will render the frame less often, but with a light-running indie game with no vsync and a much lighter use of CPU, the GPU might attempt to render 200 frames a second and overheat.

A component running at full speed shouldn’t be having problems as they’re generally designed to do it and for extensive periods of time. If it is that’s typically a sign that the component is faulty.

You can also have a problem that never appears while running them because it requires a specific feature of the hardware that almost no one uses and it just happens to be partially faulty in yours. A fun example of this is the rounding error in early Intel Pentiums. It required a very specific math problem.

https://en.wikipedia.org/wiki/Pentium_FDIV_bug

It shouldn’t, but it does, I made the same argument against XFX / AMD and they told me it’s normal (or, actually XFX was blaming AMD and AMD was telling me to wait for a new driver because I guess they thought it was a driver issue and the issue was never fixed), judging from what I hear from other people and on the internet, GPUs failing when pushed is expected these days, especially in countries that don’t have laws that make it easy for consumers to return stuff.

We must be visiting very different parts of the Internet. It might also be an AMD thing but I wouldn’t know about that as I’ve never bought an AMD card. My last Radeon was an ATI Radeon X850 XT. :stuck_out_tongue:

1 Like

XFX and AMD have seen this and told me it was normal and to wait for a new driver:
https://www.youtube.com/watch?v=TqUA6SBHGq4

(computer freezes at 0:18 and then blue screens at 2:16, this was repeatable)

There are lots of gaming PCs out there that are not 100% stable, but users claim are 100% stable. For example, I have heard users say their systems are completely stable, but then they list all of the games that periodically lockup or crash. If somebody’s computer reboots while playing a game, then that computer is not 100% stable.

Some motherboards have very aggressive settings that lead to slightly unstable results in demanding games. Unreal Engine developers have been complaining about the latest Intel CPUs and motherboards recently. I have an Intel I9-13900K system that would crash in PUBG every few hours until I turned some settings down in the BIOS. This PC passed all of the standard CPU and memory tests. Other users with a similar situation would have simply blamed PUBG, and then assumed their PC was completely stable.

Regardless of the game engine used, it is generally a good idea to give players options to self limit the framerate to prevent excessive load on their systems. For example, give players an option to enable/disable v-sync and an option to cap the framerate. Some users will appreciate the option to cap the framerate as a way to avoid coil whine in their graphics card.

1 Like

I had a Pentium 60 with that bug back then. I called Intel and they had me download a tool to run to confirm my CPU had the bug. I talked to Intel in the afternoon, and I received a new CPU the next morning. Intel handled that really well.

1 Like

Random rebooting problem is related to hardware problem, update your graphic card, RAM.

My stationary sometimes (though very rarely) reboot/powercycles… while idle, which is really odd. Seems completely unrelated to powerdraw or heat.

This is frustrating and tricky to solve since as many others have alluded to it is most likely some specific hardware issue. Added to that I’d argue customers in general are also the most unreliable people to get reports, from outright lies to misunderstandings, though mostly due to not really understanding how things work.

If I were to guess i’d echo what someone said above, that their system is likely unstable but more demanding games don’t exhibit the problem precisely because they are more demanding and thus don’t lead to a situation where a component can overheat as its limited in some way. Of course that assumes your game while it may be complex isn’t as demanding and could with unlimited rendering actually use more power ( e.g. v-ysnc disabled and no frame limiter for example)

I would suggest your next step is to investigate using Window logs to debug the actual cause of the reboot via the event viewer. You can view and save these logs ( ensure to include the display information in appropriate language) to locate the issue. You want to ensure you get events for both Application and System. I’ve not done this with someone logs, but I assume you can load them into the eventviewer on your system.

I would suggest they should retrieve these logs as soon as possible after restarting the system to minimize the amount of events you have to look through. In fact i’d go as far to insist that they shutdown the machine, start machine, immediately run your game and ONLY your game. Play it until it restarts machine, then immediately grab and save the event logs.

Having not had to do this for someone else, I would also suggest you do some googling on how to do this, especially as you’ll have to explain it to the user and then learn how to understand the events yourself. It may well be worth checking your own event log to familiar yourself with typical events from shutting down and restarting your system.

Good Luck.

1 Like