Help me diagnose random PC reboot playing Fallout 3

[blackrain]

My specs:
Phenom X3 720 BE
Biostar 790GX
Mushkin 4GB (2X2) 6400 ram
BFG LS 550
Silverstone LC13B-E case
(no video card yet - its on RMA)

All new as of May 11, so I have some time left for an RMA.



Initially, I enabled the 4th core and ran Prime overnight. No problem. Installed Fallout 3, and the PC would reboot in the middle of the game after 2-5 minutes of gameplay. On a hunch, I disabled the 4th core and the game played a lot longer after that. I checked this again (reenabled the 4th core), and had the same result. So, I guess the 4th core isn't stable despite the Prime test. In any case, with the 4th core disabled, I can play Fallout 3 for about 30 minutes to an hour before my PC reboots in the middle of the game again. I cannot consistently play Fallout without a PC reboot.

Is there any hardware that I can rule out? For example, can I rule out the motherboard? How can I test to make sure that my hardware is okay. Perhaps this is just a software issue?? Although, would a software problem cause the PC to reboot?? Wouldn't the software crash to the desktop if it was a problem with the game? I still have some time to RMA and would like to identify the problem asap.

By the way, after the PC rebooted yesterday, the CPU temp read 39 degrees.

No overclocking as of yet.

 

[blackrain]

By the way, I have onboard video. Could it be that it is overheating?

 

[WildW]

Unfortunately Fallout 3 is the most unstable game I've come across in a long time. There are huge threads about various crashes, mostly it seems caused by graphics. I went through long trial-and-error processes to get it working okay on my system.

Surprised it runs at all using on-board graphics, 790GX or not. I had more luck with ATI graphics than Nvidia, and more luck with Vista than with XP. . . but still random crashes every so often, maybe an hour, maybe less. It seems to be very finickey about the hardware it runs on. . . so the only advice I can give is wait and see what its like with your video card when you get it back.

 

[CoinOperatedBoy]

Have you tried any other games? If you have some other demanding software, run it. Does it bluescreen or just reboot? Make sure you have "automatic reboot on bluescreen" disabled in system properties. Also, the temperature after a reboot is not in any way indicative of the temperature at the time of the crash; the temp will drop significantly in a matter of seconds. You should monitor temps while the game is still running.

Overall, sounds like a power problem to me, but it could be heat. If you have another power supply to test with, I'd try that. Use a multimeter to test your voltages if possible; see the power supplies subforum for instructions if you need them.

Fallout 3 is also kind of a buggy piece of shit, so it could just be that.

 

[blackrain]

For such a cool game, it is disappointing that it is so unstable. I did google this specific problem before posting, and I did find a number of instances of PC/computer reboot with Fallout 3, but not as many reports as I had hoped. No solution was acknowledged. At this point, I just want piece of mind that my hardware is good (since it is a new build).

What other games would you suggest? I just got Mass Effect, but have not installed it yet. Would that be a good test?

Should I run Prime again? If so, how long and on what settings?

There is no blue screen. Where in "system properties" is the option to disable automatic reboot?

How do I monitor temperatures while the game is running??

 

[CoinOperatedBoy]

Right-click on My Computer > Properties > Advanced tab > Startup and Recovery button > Uncheck "Automatically restart" in the System failure box.

If you don't do this, a bluescreen may be too quick for you to even notice because the system will instantly reboot.

Any other modern game, Mass Effect included, is at least worth trying to tax the system in a similar way to Fallout 3. It's not the same as running stress test software. But you might also try running OCCT's Linpack test for a while.

Use a program like CoreTemp to monitor your temperatures. It has logging features. Play a game for a while, then alt-tab out and check your temps, or quit and check the log.

 

[blackrain]

If there is a reboot, will the temp at the time of reboot be saved/logged using CoreTemp? Or should I should play for a half hour, and then check?

regarding running OCCT, what is "a while"? 2-3 hours? overnight?

 

[zerogear]

There will be a log, though I think you need to enable it. Did you check Windows event logs yet?

 

[CoinOperatedBoy]

Originally posted by: blackrain
If there is a reboot, will the temp at the time of reboot be saved/logged using CoreTemp? Or should I should play for a half hour, and then check?

regarding running OCCT, what is "a while"? 2-3 hours? overnight?

The temp should be continuously logged on some interval. Your temps should rise significantly almost as soon as you start the game. I would play for 10 minutes or so, then check. Then play until a crash and examine the log.

As for OCCT, it's kind of up to you. You usually don't need to run it as long as you might let Prime95 go because it tends to produce errors sooner if they're going to happen at all. For an initial test, 1 hour might be fine. If you really want to stress test for stability, I would say overnight is reasonable. Be aware though that your temperatures may be up to 10C higher than with Prime95, so use caution.