GPU temp readings are too low to be believable

GhostGamer

Honorable
Jan 4, 2013
17
0
10,510
Hey everyone,

I have an R9 280 Royalking from Club3D, and i have overclocked the card to 1155 mhz core clock, 1263 mV and 1575 mhz Memory clock. The fan is running at 80%, rpm is about 3200. Idling i get anywhere from 24 to 36 c temps. Under load my temps are 50-63 c. The problem is that whenever i touch the GPU heatpipes when playing a game they feel awfully hot. Like really hot. If i put my finger on it for longer than 3-5 seconds it actually starts to hurt. GPU-Z is reporting the same temps. Anyways, today when playing the card throttled down to a whopping 300 mhz core clock. So something is wrong if the readings say that the GPU is at 50-60 c. I have never heard of throttling under 80c. I suspect that the thermometer in the GPU has gone bad. What do you guys think is going on? Is/are the heatsink/heatpipes supposed to feel this hot at 50-60 c?
 
Solution
55°C = 131°F, yes that's rather hot to the touch. As far as the throttling goes, I'd back off on the OC and reset default clocks. It's entirely possible your OC is demanding too much current to maintain the voltage settings, so during an intense gaming session is spiking the temps. If you have a lower quality psu, that has a bad ripple output and lower than 12v output, this will only make things worse.

The temps as is are pretty much exactly the same as my 660ti, but whereas I'm running stock clocks currently, you are running an already known warm running AMD card with OC and 80% fan, so that balances somewhat.

Basically I'd be looking for another explanation for the throttling, and I'd start with the psu outputs read by multimeter...

Karadjgne

Titan
Ambassador
55°C = 131°F, yes that's rather hot to the touch. As far as the throttling goes, I'd back off on the OC and reset default clocks. It's entirely possible your OC is demanding too much current to maintain the voltage settings, so during an intense gaming session is spiking the temps. If you have a lower quality psu, that has a bad ripple output and lower than 12v output, this will only make things worse.

The temps as is are pretty much exactly the same as my 660ti, but whereas I'm running stock clocks currently, you are running an already known warm running AMD card with OC and 80% fan, so that balances somewhat.

Basically I'd be looking for another explanation for the throttling, and I'd start with the psu outputs read by multimeter, not software, bios if you must.

One thing to remember, your gpu Vram temps are different entirely to the gpu cpu temps, which are what's normally read by gpu-z etc, so while the gpu may read normal temps, the vram may be off the charts.
 
Solution

Unclejimbo

Reputable
Nov 11, 2014
3
0
4,510
I'm currently having the same problem i didnt overclock my gtx 670 but the temps are Low yet touching the heatpipes would cause burned fingers.

Just checked FURMARK directly when the test starts its throttling from core:914mhz (throttling: - 364mhz). After less then 1 minut the heatpipes are smoking hot, yet the temp's in FURMARK seems to be normal.

Did you find the problem allready ?