Sign in with
Sign up | Sign in
Your question
Solved

PC shut down while gaming

Last response: in Overclocking
Share
October 31, 2011 5:55:20 PM

New build:
MSI P67A-GD53 (B3)
Intel Core i5-2500K Sandy Bridge 3.3GHz
EVGA SuperClocked 01G-P3-1563-AR GeForce GTX 560 Ti (Fermi) 1GB
CORSAIR Vengeance 8GB (2 x 4GB) 240-Pin DDR3 SDRAM DDR3 1600
COOLER MASTER Storm Scout SGC-2000-KKN1-GP Black Steel / Plastic ATX Mid Tower Computer Case
Western Digital 1TB hard drive 7200 rpm
Corsair 500w PSU
1 DVD drive
COOLER MASTER Hyper 212 Plus cpu cooler
3 120mm fans on back top and front of case

Built this rig yesterday and temps under load were about 35c idle and 58c 100% load for cpu. I used oc genie to get cpu to 4.2z and it idles at about 38C and load 68c. I was playing a CPU intensive game that uses about 60-75% of my cpu for about 6 hours when my computer just shut down. The temps in that game only get to maybe 60c so im wondering if its my psu. Also my GPU idles at about 35c but while playing bf3 at 100% load it gets up to 80-85C is this fine or should i be worried about it?

More about : shut gaming

October 31, 2011 7:09:23 PM

80-85 in BF3? That's totally normal, however, I wouldn't feel safe with a 500W PSU on that rig to be honest with you. It's the Corsair Builder Series correct?
m
0
l

Best solution

a b à CPUs
a b K Overclocking
October 31, 2011 8:06:40 PM

Edit: Sorry, I misread your cpu temp as a gpu temp. but establishing max temps under known load remains important.

You say your GPU runs 68C "under load", and then say it runs 80C-85C when running BF3. This makes me want to ask what known loads did you use to stress test your build, and your OC, and for how long?

What I'm worried about is a build up of heat inside your case over time that gradually raises the ambient temperature inside to the point of shutting down a component. You may not have realized that heat exceeded max when the failure occurred.

The purpose of stress-testing is to assure temp and operational ability under max load, which apparently wasn't done based on your temp reports.

Let's try some known loads and see if we can't replicate the failure:

Download the latest CPUID's Hardware Monitor, Prime95, and Furmark from this site:

http://forums.tweaktown.com/gigabyte/30530-latest-overc...

Run Prime95 stress test for 10 minutes (not too worried about cpu temp atm) and let us know the max temp reached, AND whether the cpu was "hovering" at a specific temperature by the time the test was over. This would indicate temp control is good. If P95 fails, there's probably a memory issue.

Run Furmark Torture Test for 10 minutes. Please report same info as for cpu.

Run both together for at least 5 minutes. Temps should hover at same or slightly higher than the setpoints established above.

If at any time, a cpu core temp hits 80C or gpu temp reaches 90C, stop the test and resolve the heat issue.

(I'd also run a P95 overnight to validate stability of the overclock.)

If these temps show "nothing's wrong", I'd resume playing games making sure that HWM was logging temps so I could go back and find what the temp was the next time it fails.

The problem could be your psu is not delivering clean power after long runs (ie its defective), but the 500W capacity is not an issue in itself.
Share
Related resources
October 31, 2011 8:18:02 PM

Twoboxer said:
You say your GPU runs 68C "under load", and then say it runs 80C-85C when running BF3. This makes me want to ask what known loads did you use to stress test your build, and your OC, and for how long?

How did I miss that....? :pfff: 

But yeah, try boxer's test and see if your OC is stable.
Post the results once your done!
m
0
l
a b à CPUs
a b K Overclocking
October 31, 2011 8:20:49 PM

ROFL, Sushi . . . only reason you "missed it" is that you haven't (yet) been burned as often as I have lol.
m
0
l
October 31, 2011 8:36:22 PM

I'm looking forward to being burned.
m
0
l
October 31, 2011 8:48:31 PM

sorry u misunderstood me lol all of that was for CPU other then last sentence which was my GPU. All of these load temps were with prime95 on for about 30min to 1hr with core temp and speedfan. I havent seen my cpu go above 70C which makes me believe it couldn't of been my cpu that shut it down. I guess after long hours of gaming it can get higher temps then 1hr of prime95?
m
0
l
October 31, 2011 8:50:32 PM

and yes its a corsair 500w psu about 2-3 years old (used in last build) never had a problem with it.
m
0
l
a b à CPUs
a b K Overclocking
October 31, 2011 8:57:25 PM

Assuming you ran Prime95 for >30 min after you OC'd, no need to duplicate that.

But P95 doesn't touch your gpu. Please do the Furmark and P95+Furmark tests above. You might see the gpu temp fail to peak and "hold" under one or the other.

And please substitute CPUID's Hardware Monitor for Speedfan to report gpu temps here. Makes things easier on us because we trust that utility. CoreTemp is OK and trusted for cpu, but HWM does do both.

m
0
l
October 31, 2011 9:02:23 PM

Ok ill try that but for GPU temps i was using evga precision. Can my CPU get hotter if ran for a long time, and also does it make sense that it could be my psu that shut it down?
m
0
l
a b à CPUs
a b K Overclocking
October 31, 2011 9:37:53 PM

Yes, your cpu and gpu temps can get hotter if run for a long time. It depends on how well your case exchanges air.

Yes, your psu could have shut it down. It might be that your 2-3 yr old psu is degrading, and this is the first sign of its impending failure. As you can understand, PSUs do not last forever.

Yes, it does makes sense that your PSU could be the thing that shut down your system. The ONLY way to test that is to swap in a known-working psu.

However, we will not recommend that until we exhaust internal heat issues on other parts . . . these we CAN test for, and you haven't done that yet.

As requested, for this gpu temp test please use CPUID's Hardware Monitor. CPUID's GPU-z is an acceptable alternative. Frankly, I virtually ignore all other reports because there's no way for me to calibrate them.

What we're looking for is: does your gpu reach and hold an acceptable temperature when running at what we know would be full load, ie, both P95 and Furmark.
m
0
l
October 31, 2011 10:13:11 PM

So i ran p95 and furmark at the same time for 15min.

CPU temps average was 69C max 73C
GPU temps average was 87C max 89C

Something weird happend tho at 15min of furmark my screen froze and my mouse and keyboard shut off and my fans noise lowered but everything in the computer was still running, I ended up having to hard shut down.

I have never seen a lockup where the leds on my mouse and keyboard shut off. Do u think i should try running the benchmarks again?
m
0
l
October 31, 2011 10:52:18 PM

I'm not completely 100% sure, but it sounds like your GPU is overheating. My GPU only gets up to about 75-80C During Furmark.
m
0
l
November 1, 2011 12:13:40 AM

So last time for Furmark i bumped the resolution up to my native which is 1920x1080 and that produced those high temps.

This time i didnt touch any settings and ran both p95 and furmark again for 45min.

CPU temp hovered at about 67-68C with 72C max
GPU temp hovered at about 81-83C with 85C max

My GPU fan speed while running furmark was only running on 51% which was the default and i read somewhere that the default is terrible and to make some profiles in evga precision to ramp up the fan speed to 75-85% during load.
m
0
l
November 1, 2011 12:32:48 AM

Now it sounds correct. I'm pretty sure its your PSU now, but, don't depend on my decision since I'm a total dunce with GPUs. :kaola: 
m
0
l
November 1, 2011 12:37:44 AM

So i went into evga precision and turned on the software automatic fan control and it ramps up the fan speed 10% over 50% for every 10C over 50C

so now during furmark my card is chillin on 74C with max 75C :) )) (ill deal with the increase in noise)

DERP! Antec Earthwatts 500W PSU, not corsair. sorry XD

Also if it is my PSU that shut down my system before should i be concerned for my other hardware or will it just be the psu that goes out?
m
0
l
November 1, 2011 12:52:34 AM

WTF at 19min of furmark it did the same lockup as before. GPU temp was at 74C CPU temp was at 68-71C across all cores. Everything seems fine.....

I dont understand, I ran it for 45min before with higher temps on the GPU and it didnt freeze. Even if it is the PSU what is making it use more power and freeze?
m
0
l
November 1, 2011 12:53:12 AM

You should be concerned and should get a replacement ASAP. From my past experiences, a faulty PSU is like a ticking time bomb. One or two of my friends have actually had small fires from sparks and wires. One of them had his hard-drive taken out, thank god it was already suckish. ;P

EDIT:
Please, use the edit button if your going to post quickly in succession.
m
0
l
a b à CPUs
a b K Overclocking
November 1, 2011 12:58:17 AM

One last thing you could try . . . clear cmos, and un-OC both the cpu and gpu by loading BIOS defaults. Uninstall all utilities that can overclock. If you get the same failure, then swap in a new PSU
m
0
l
November 1, 2011 1:44:46 AM

Twoboxer said:
One last thing you could try . . . clear cmos, and un-OC both the cpu and gpu by loading BIOS defaults. Uninstall all utilities that can overclock. If you get the same failure, then swap in a new PSU

Clear-CMOS is your friend.

Remove your motherboard's battery also, wait for a few seconds, put it back in, then CLR-CMOS.
m
0
l
November 1, 2011 3:32:35 AM

So i "unoverclocked" my CPU by pressing my oc genie button again and rebooting. i benchmarked for 1h 10m and everything was fine so i pressed it again to overclock back to 4.2ghz and am benchmarking again... been 45min and it hasnt crashed yet.

Seems fine with max CPU temp at 71C and GPU temp at 75C. I dont know if its a good or bad thing that i cant recreate the failure XD
m
0
l
a b à CPUs
a b K Overclocking
November 1, 2011 5:18:59 AM

If it's still there, not to worry - it will return :) 
m
0
l
November 1, 2011 5:43:41 AM

Best answer selected by bobman21.
m
0
l
November 1, 2011 11:29:09 PM

So I ran a p95 test overnight and when i checked it in the morning it had crashed the same way as before so it def isnt my GPU.

My OC Genie overclocks my CPU to 4.2 ghz with 1.38vcore which seems pretty high voltage for 4.2 ghz. Could that be what is causing the crash? If I do it manually and set the voltage a tad lower do you think that would do anything?
m
0
l
a b à CPUs
a b K Overclocking
November 2, 2011 2:39:31 AM

First step is clearly going back to stock settings to establish whether or not it *is* your OC that's the problem.

So disable/uninstall any program that can OC anything from Windows. Then clear CMOS and load defaults. Re-run the P95 tests. If that runs overnight, you can go back and play with the OC. If it doesn't run P95 overnight, we've got a different problem to solve.
m
0
l
November 2, 2011 9:06:37 PM

So I unOCed the CPU by just pressing the genie and p95 has ran for about 15 hours lol and its fine. I guess it was the overclock that was causing it. Does this mean i shouldnt overclock this chip at all or maybe the genie was putting too much power when it isn't needed and thats causing it. Do you think if I do it manually and set voltages more accordingly it will be more stable?

Also this freezing that happens after some time is not the same as the crash the first night i was playing on this comp. The first time the comp actually shut down instead of freezing, but it hasn't done that since.
m
0
l
a b à CPUs
a b K Overclocking
November 3, 2011 3:14:22 AM

You are lucky that "just pressing the OC button" worked well enough.

If one push gave you a bad OC, are you sure another push gave you a clean stock PC? I don't think so. There are good reasons we ask you to do things a certain way.

Clear CMOS and load defaults. After running a few days and being dead certain it was stock all around, press the button and try the OC again. Maybe it will work if it starts with a clean stock PC.
m
0
l
!