Please help: Inexplicable restarts

[gonneman]

Hello Tom's Hardware community!

My home built system is suffering from really annoying restarts since last Saturday (25 Aug).

They happen while playing Diablo 3 and Guild Wars 2. They system will, after a random amount of time in the game, just power down and then power up again. I say power down because it really feels as if someone pulled the plug. All the fans and such stop and then start up again.

Since this happens in more than one game, I don't think it is the game's fault. If anyone could give me any insight on this I'd be very grateful. To give you as much information as possible I'd like to first list my specs and then everything I've tried in order to diagnose the problem.

CPU: Intel Core i7-2600K Processor
4 cores that split into 8 clocked at 3.691 GHz
GPU: NVIDIA GeForce GTX 570
with 1280 MB ram, clocked at 750 MHz
Ram: 8GB
Mainboard: ASRock P67 Pro3
OS: Dual boot 64 bit Windows 7, 64 bit OpenSuse
No over clocking or anything like that

What I've done so far:
Last Sat 25 Aug the CPU fan came loose, so I renewed the thermal paste and refastened it. My first thought was that the
CPU might be over heating. But after refastening the CPU, fan temperatures now plateau at about 60 C.
The GPU goes up to about 70 C but is stable there. That isn't too high, right?

I thought maybe the CPU is damaged, so I let prime95 run to stress test the CPU for about 20 min without any problems.

I ran memtest86 to test my ram, also without any problems.

I tried 3DMark11 to test my graphics card. The crashes happen while gaming after all. It runs through an entire benchmark without problems.
I haven't found a benchmark yet that will run for an hour or so to really simulate gaming.

I reinstalled Windows just to make sure it wasn't some virus of driver problem. But Guild Wars 2 and Diablo 3 still lead to restarts.

Nvidia geforce drivers 301.42 (latest stable), 306.02 (latest beta), 296.10 (2nd most recent stable) all lead to the same restarts.

The fact that everything powers down might be due to loose connections after fiddling around in the PC refastening the fan. I recheck all the connections and pluged the PC directly into the wall instead of the power strip. Still no improvement. The power supply isn't getting hot, so I don't think it is that either.

I can't reproduce the restarts in way other than playing games. The time it takes for the restart to occur does not seem to follow any pattern. It can happen after a minute or after an hour. The restart happens to suddenly for the Windows event log to register anything.

So now I have run out of possibilities I can think of. Does anyone have any recommendations? This really is very frustrating.

 

[lourendo]

what PSU wattage do you have and is a reliable manufacture? you might be running into a low power situation when the pc gets hot and the fans speed up etc. i would run a system temp logger to record your temps at the point of the crash. may be a heat issue with northbridge or hdd etc. also, what kind of case fans do you have? adequate airflow?

do you do any overclocking?

 

[gonneman]

No overclocking.

My PSU is 650W and it is made by Scythe, which I believe is a fairly well known brand.
I run speed fan and all the temperatures are fine CPU plateaus at 60C GPU at 70C. The rest is about 30-40C so quite cool. I have 5 case fans not including the PSU or fans on CPU and GPU.
The air flow ought to be fine since the PC has worked without any problems for about a year and I haven't changed any of the hardware.

Any other ideas?

 

[lourendo]

try opening blender and rendering an insane number of polygons, as in a several hour render time. if it does not crash it may be bugs in the game or still hardware issue. if not it most likely is a software issue. may be viruses etc. if you can find one run a process logger and see if any strange processes or services are running during crashes. might just be a bug in the driver or something.

 

[gonneman]

In a rather bold and reckless move I tried turning of the overheating protection in the bios. So the system wouldn't automatically restart if things get hot. To be safe I set up alarms in Speedfan instead.

The PC still crashes. I'm really wondering what can cause this. There is no blue screen or any thing (I've set the system to wait on blue screen, not restart). The PC just powers down and then powers up again.

Failing drivers and the likes should not produce this behavior right?
What can cause the system to restart like that?

 

[lourendo]

insufficient power among other things. the gpu could still be overheating rather than the cpu. run memtest overnight just to rule that out.

 

[yogy]

try furmark as well

 

[gonneman]

I did memtest for about 5 hours without finding anything and also did the 15/20 min benchmark in furmark. No errors occured. I'm now sending cpu, gpu and mainboard back to the shop where I bought them for testing.
I hope they find something....

 

[lourendo]

thats certainly a thourough way to get it diagnosed.

 

[Penguin_13]

I did memtest for about 5 hours without finding anything and also did the 15/20 min benchmark in furmark. No errors occured. I'm now sending cpu, gpu and mainboard back to the shop where I bought them for testing.
I hope they find something....

Did they find anything wrong? We are having an identical problem, random reboots only while playing Diablo 3 or Guild Wars 2.

OS Name Microsoft Windows 7 Home Premium
Version 6.1.7601 Service Pack 1
System Type x64-based PC
Processor Intel(R) Core(TM) i5-2550K CPU @ 3.40GHz, 3701 Mhz, 4 Core(s), 4 Logical Processor(s)
BIOS Version/Date Award Software International, Inc. FB, 10/12/2011
SMBIOS Version 2.4 Windows Directory
Physical Memory (RAM) 8.00 GB
Graphics Card NVIDIA GeForce GTX 560

Not sure what brand etc. our mainboard is.

P.S> ran the same tests on memory, graphics card, etc, even bought a power supply tester and did that...everything came up fine.

 

[lourendo]

I think its your GPU. try updating your drivers and tweaking game settings.

run it with the new drivers and see how it does.


run on low settings and see if it still crashes. if not...

use some binary exclusion to rule out settings that may trigger the crashes

 

[gonneman]

OK I've finally managed to fix the problem (a few weeks ago acutally, but I didn't want to celebrate before I was sure).

So the shop did a 12 hour load test on the CPU, GPU and mainboard, but could not find anything.
However after reassembling my PC the crashes became less frequent, but they did still occur. So to be sure it really wasn't the RAM, I tried running the system on just one bar of RAM and swapping when crashes occurred. Both bars lead to the same kinds of crashes.

As a final desperate measure I bought myself a new PSU with higher wattage and greater efficiency. Which finally fixed the problem. I haven't had any crashes for a month! I have no idea why my PSU suddenly developed these problems after just a year use, but at least things are working again.

Thanks lourendo and yogy for your advice.

 

[gonneman]

Best answer selected by gonneman.