GPU temp high and unexpected shut downs

[Inceington]

Hi,

Specs below, I had a couple of unexpected shutdowns during gaming and googled a bit to look at reasons. The most likely seemed to be heat so I looked at the GPU temp in the AMD driver thing and the 5770 idles at approx. 62 and under load (before shutting down) it gets to over 100 degrees. From what I can tell this isn't good. I've been running this card about a year but it is older than that as I got it from a friend. The other option I read about was a PSU fault. What I couldn't find was if there was any way to tell which one was the problem before replacing one. Any help would be really appreciated and any general advice on the whole problem too. I'm pretty new to all this so I could be barking up the wrong tree!

Thanks in advance

Motherboard: MSI 990XA-GD55
CPU: AMD FX-6 6100 3.3Ghz
Memory: 8GB G-Skill DDR3 1600mhz Sniper CL9
Storage: Seagate 500GB 7200RPM 16MB Cache
Graphics card(s): XFX HD 5770
PSU: Coolermaster GX 650w
Case: Coolermaster Centurion 5 II
Operating System: Windows 7 64bit
Monitor(s): LG LCD 1080p 23"

[technodean]

First port of call would be to make sure the fan is working as it should, i.e, keep an eye on it as it could be intermittent/ failing and that the cooling is not covered in dust blocking the airflow. Also you could download msi afterburner and use the monitoring during gaming which if you have set like I have then will show your fan speed in % and your temps. when or if it gets high quickly look at the gpu's fan. Alternatively, you could always set your fan speed in ccc to around 60% etc where it is very audible so you would hear if it was failing.

[Inceington]

Thanks, I tried setting the fan to 100% in ccc and stressing. The card got to 101 degrees relatively quickly (a minute or two) and the fan didn't sound like it stopped working at any time.

[Phage]

Probably full of dust or the cooler is no longer fully n contact with the GPU. You need to take it apart, clean and reseat the cooler.
Or, it's time to trade it in.

[Inceington]

Some new information.

I'd decided that the GPU was the problem, seeing as there wasn't dust anywhere really and the fans were working etc. So this morning I played a video in VLC, just to see if that would have much of an effect. So i was checking GPU temps and they were fine but I guess I got lax and about 15mins in the PC shut off and wouldn't turn back on again. I put my hand in the case and it did feel pretty hot, but it was the CPU cooler that was hottest, not the area around the GPU. When I got the PC back on, after it had cooled I assume, the CPU temp was 57 degrees I think. Anyhoo, I've been running core temp and repeated the experiment but both my cpu and my gpu temps are stable, the first at about 40 or less and the latter at about 62. I tried gaming again and the GPU still shoots up, CPU temp did rise but only to 55-60. Do I need to worry about my CPU? Or does this information make a different diagnosis likely? I just don't want to waste money on a GPU if thats not the problem!

[csgohan4]

You could try Prime 95 your CPU WITHOUT the GPU as a control test and if it runs fine for say an hour then you should be fine and it's most likely your GPU. Btw did you Overclock your system i.e CPU or GPU?

[Mr_Jon]

which way is your psu mounted?

if it's facing up, it could be sucking in all that hot air from the gpu and case, which could cause it to shut off. your psu is rated for 50 degrees though, it's a long shot.

this review doesn't paint the psu in a good light though.

[Inceington]

The PSU is facing downwards, and yeah at the time I thought it was a decent choice but since I've read a couple of reviews like that. Do you think that could be my proble

@csgohan: I haven't OC'd either and I have no onboard graphics so I kinda can't test things without the gpu lol

[Inceington]

Got CPU usage to 45% and have max cpu temp of 63 degrees btw. Not really sure what to make of that number

[Mr_Jon]

The PSU is facing downwards, and yeah at the time I thought it was a decent choice but since I've read a couple of reviews like that. Do you think that could be my proble

i'd say it's an outside chance, but only if the gpu is ruled out. it's likely the gpu, idling at 62 sounds very high.

[Inceington]

Is Prime95 dangerous in any way?

[Inceington]

I tried Prime95 with an In-place large FFTs for 6 tests. I only got through 3 before I manually stopped it because the core temp read as 86. This only took about 5mins I didn't get any errors or anything but that seems really high to me. Is that bad news for my CPU?

BTW, just wanted to say thanks to everyone whos helping out here, its really appreciated

[kalniel]

86 is not too high for a CPU, but it is pretty warm, warmer than I would expect if your system cooling was adequate.

The centurion 5 is an OK case.. but either you've got no cooling at all or it's badly misconfigured, or you've got some component in your case heating the insides up faster than the heat can be taken out.

[csgohan4]

how many case fans do you have and which are blowing in or out of the case

[csgohan4]

Is Prime95 dangerous in any way?

Prime 95 is not dangerous, it is a CPU stress test to test stability of overclocks which is the commonest use. It is dangerous if you use it incorrectly.

[Inceington]

how many case fans do you have and which are blowing in or out of the case

one 140mm blowing in at the front bottom and one 120mm blowing out at the top of the back.