My pc shuts down (no reboot) while gaming

[matrix1700]

Everytime I play games, such as Bionic commando, Anno 1404, COD 5, my computer shuts down after 5-10 minutes of gameplay. It's like someone pulls the plug. After that, I can turn on the computer without facing any issues.

At first I thought my cpu was the problem, cause apparently my q9300 ran (idle) at 60(according to speedfan) degrees celcius. So I cleaned my case and my cpu cooler(scythe mugen infinity), removed the dust and the temps dropped with almost 10 degrees. However, the shutdowns still occur.

I think my cpu cooler isn't working properly or might it just be my psu that is failing constantly? I really need to get this pc up and running again, for there is much gaming to be done.

Any ideas?
Many thanks in advance.

*update*: I read that increasing my cpu temp limit in the bios might help. It's currently on 45. Is this a good idea and what would be a suitable temperature to put the limit to?

[nightkhaos]

Download OCCT (Overclocking Checking Tool) and check what your load temps are. 50 degrees idle does seem a little high (unless, are you still using the stock cooler?).

If they check out okay, we shall look into your PSU, but given your hardware I hardly see why that PSU would fail.

[Salazaar]

Is it only when you're gaming or does it happen when you're doing other CPU intensive tasks?

[loriis]

Also check out the northbridge temps under load (if you can), I had a board with a faulty nvidia chipset recently, and it would overheat and act in exactly the way you describe.

[matrix1700]

@ nightkaos:
Just did the test. The results are as follows:

- Cpu 1 = 59 under load
- Cpu 2= 64 under load
- Cpu 3 = 67 under load
- Cpu 4 = 61 under load

Acceptable or risky?

@ Salazaar:
Only when im gaming. The other tasks I run aren't as demanding as the games I play.

@ loriis: 33-35 celcius

Thanks for the quick reply guys!

[simonw]

Those cpu temps look fine to me.

What about gpu temps? You can measure the temps with GPU-Z, perhaps do a gpu stress test using FurMark or similar.

[nightkhaos]

Those cpu temps look fine to me.

What about gpu temps? You can measure the temps with GPU-Z, perhaps do a gpu stress test using FurMark or similar.

I agree. GPUZ check now.

[Salazaar]

@ Salazaar:
Only when im gaming. The other tasks I run aren't as demanding as the games I play.

In which case I'm inclined to agree with the others, take a look at the GPU rather than the CPU. Though it would be worth downloading some sort of CPU stress test/benchmarking utility just to rule out the CPU.

[nightkhaos]

In which case I'm inclined to agree with the others, take a look at the GPU rather than the CPU. Though it would be worth downloading some sort of CPU stress test/benchmarking utility just to rule out the CPU.

*hemhem* Why do you think I told him to download OCCT?

[matrix1700]

Yep,

The gpu was causing the shutdowns, temps came near 120 celcius, and then Thermal protection started kicking in I guess.

I already got the fanspeed on 100%,

Should I buy another card?

[simonw]

Ah, finally getting to the root of the problem. If the gpu card is fine for what you use it for, then I would probably do a couple of things first:

1. With a 120c temp, that would suggest the hsf may not be installed correctly. Try reinstalling it and apply a new layer of thermal paste to check if that helps matters.

2. If you are still getting high temps, perhaps invest in an aftermarket cooler, definitely cheaper than a brand new card.

[holy god of nil]

If it's new I'd look at rma'ing with temps that high, surely must be a fault?

[Salazaar]

*hemhem* Why do you think I told him to download OCCT?

No need to get defensive mate, no one here's got a monopoly on trying to be helpful.

[watercooled]

If it's new I'd look at rma'ing with temps that high, surely must be a fault?

Yeah I definitely would, provided it's still covered my warranty. What card do you have?

[AntiTalent]

Ah, finally getting to the root of the problem. If the gpu card is fine for what you use it for, then I would probably do a couple of things first:

1. With a 120c temp, that would suggest the hsf may not be installed correctly. Try reinstalling it and apply a new layer of thermal paste to check if that helps matters.

2. If you are still getting high temps, perhaps invest in an aftermarket cooler, definitely cheaper than a brand new card.

^I agree.

Clean out any dust from the cooler, and I'd definitely reseat it. Some of the factory applied TIM is way too thick and shoddily done. I had an XFX card with loads caked on. A reseat helped drop the temps a fair way, but in the end I went for an aftermarket cooler.

Of course you could just RMA it, but a clean out of any dust will cost you nothing, and might well help.