Computer restarting

[Jan Zich]

Hello,

I'm looking for an advice with a crashing / restarting computer.

Few months back I bought and built a new computer. The problem is that it occasionally restarts itself. I use the computer about an hour or two every day every evening and little bit more over weekends. This occasional restarts happens about once a week, sometimes even less. So far, I think I've experienced four or five restarts. The main problem is that I'm not able to associate it with any particular activity.

To be more specific, it is really just a complete shutdown and restart, no blue screen. It looks like a short power issue, but I'm guessing that it's really the motherboard detecting some issue and shutting down to prevent more damage (or something like that). As I said above, I can't really tie it to any particular activity. I think in about half of the cases it happen when I played a computer game (Age of Empires Online to be precise which is annoying because there is no auto-save), but in the other cases in happened when doing something completely harmless such as typing in a text editor or watching a youtube video. Also, I played and finish Bulletstorm on the computer without any issues.

I'm guessing that if it was overheating of some sort, I would see consistent failures during certain activities. But when it happened the first time, I ran some memory, CPU and GPU stress tests for longer periods of time (not hours, but tens of minutes) watching the temperature climbing and eventually stabilizing but not crashing the computer.

Here are the components:

• Intel Core i7-2600K
• G.SKILL Ripjaws Series 8GB DDR3 1600 (2 x 4GB)
• ASUS P8P67
• EVGA GeForce GTX 570
• SilverStone Fortress FT02B
• Antec EarthWatts EA750 (750W)
• Intel SSD 320 160GB

I know, there are not fancy CPU fans or anything like that, but I'm not a hardcore gamer or overclocker, so I figured I should be fine with the stock fans. If I had to pick, I would say it's the memory most likely, but as I mentioned above, I've run memtest for several iterations with no issues. I check the motherboard manual - I suspect there is something about error detection and restarts.

I'm OK to replace any component ... only if I knew what it is.

Any advice would be greatly appreciated.

Thank you

 

[AnonymouseUser]

"Several iterations" of memtest is simply not enough testing. I've seen ten or more loops before any errors cropped up, then it was a deluge. Always run memtest overnight when checking for RAM issues.

The crashes could also be caused by drivers, so make sure everything is up to date.

You could also check your logs with Event Viewer.

 

[Jan Zich]

Thank you, I will run memtest overnight. Also, I've just reset the BIOS to its defaults to be sure (but I think there was not anything exceptional set before).

Regarding the Windows Events; I had never found anything other than:

The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

or

The previous system shutdown at 9:07:06 PM on ‎9/‎16/‎2011 was unexpected.

after the restarts.

 

[AnonymouseUser]

I guess I should say, I've also seen overnight testing with only one or two errors total. These can be caused by mobo or RAM, so you use the process of elimination to determine which is faulty.

As for the event logs, it does seem to point to a hardware issue.

 

[Zaap]

First thing I recommend after testing more obvious general causes is google your motherboard model + the problem. (IE: ASUS P8P67 random restarts) Chances are someone else has the same problem. That one is unusual and random enough it may be motherboard specific.

At just a glance I found several people experiencing the same issues, and their various solutions/attempted solutions.

In particular, read the first and last posts here and see if the solution works for you as well:


http://vip.asus.com/forum/view.aspx...d=1&model=P8P67+DELUXE&page=1&SLanguage=en-us


This the Deluxe version of the board, but since the symptoms are nearly identical, it could possibly be the same cause and common solution, IE: turn off automatic system error reboot and the load line setting in the BIOS.

 

[Jan Zich]

Thank you for the advice.

I ran again the memtest. Now overnight. It did 10 passes without any errors. Then I reset the BIOS to its default settings and for three days the computer seemed fine until now. It just rebooted again after playing Age Of Empires for about an hour. I'm not an expert, but if it was overheating, I would expect more frequent and more predictable crashes.

Also, I inspected the BIOS settings mentioned in the ASUS thread.

 

[Jan Zich]

OK, I think I’m finally getting somewhere. I’ve just been playing with Prime95 and it seems I can get the system to crash consistently within several minutes. I think when I noticed the issue the first time some time ago I did some CPU stress tests, but not this one. In any case, I think it’s a good news.

I’ve done several iterations. Sometimes it takes not even a minute, sometimes more. Sometimes, the system reboots, sometimes freezes (no blue screen, just frozen screen). Now, this may get embarrassing, but I think the CPU temperature gets out of control. My normal idle temperature is about 40-42C and when running Prime95 it quickly climbs to about 90C where it stabilizes and then freezes after time one. Interestingly enough, when the system just reboots, it happens much faster and during much lower temperature. Usually around 70C which is still presumably a not a completely unreasonably high temperature.

As I pointed before, I’m using the stock CPU fan, which (according to the Internet) should be still usable even for reasonable overclocking. I’ve also read that it could be caused by an improper installation. I’m going to open the computer and check it …

 

[Jan Zich]

Update: After seeing the temperatures, I bought:

http://www.newegg.com/Product/Produc...82E16835118074

I just installed it, and to my surprise, the problem did not go away. The idle CPU temperature is now around 30C, and when I run Prime95, it climbs to and stabilizes at 50C. But still, the computer restarts within 5 or 10 minutes. I’m considering to replace the motherboard. Does it sound like a good idea? Or the CPU?

 

[AnonymouseUser]

That's a huge temperature difference!

So is Load Line Calibration set below Extreme in the BIOS per the thread linked earlier? Can you log the system voltages so you can see if they fluctuate before the crash?

You may even have an overheating chipset on the motherboard, so check that those heatsinks are seated properly.

 

[the182guy]

Have you tried cleaning out the fans? Shouldn't have built up much dust in a few months but it will still help.

 

[Jan Zich]

It seems that my idle vcore is something like 0.96V. I did one more Prime95 test before I went to bed yesterday. It happened to run 10 minutes and did not crash, but vcore was between 1.152V and 1.184 V (the exact distinct values I saw were 1.152V, 1.160V, 1.168V, 1.176V and 1.184V) averaging at 1.1659V. Is it normal? Is this the temperature I should be watching? I will do some more tests to actually crash it. It usually takes much less than 10 minutes.

Otherwise, everything in my BIOS is set to default. The Load Line Calibration settings it set to Auto.

 

[AnonymouseUser]

Those fluctuations are probably fine since it didn't crash (tolerance is +/-5&#37, but do log them so you can see what they are just before it crashes.

Also, set Load Line Calibration to High so it doesn't automatically get set to Extreme.

 

[Jan Zich]

Thank you, I will give it a try.

I did one quick Prime95 run this morning and it rebooted within two minutes (this is the more common case). Vcore averaged at 1.165854V within +/- 2%.

 

[Jan Zich]

But reading a bit more about the “Load Line Calibration”, it seems to me it has more to do with regulating Vcore during changes of the CPU load. My computer crashes when it’s under full constant load. At least when I run Prime95 in which case the crashes are very consistent. When it crashes during the normal usage, it seems quite random (as I described in the original post) that could be potentially related to the LLC.

 

[THRiLL KiLL]

i may be off kilter here, but i would try lowering the speed of the memory.