GPU overheating to the point of throttling

[Juliuz]

My GPU is overheating, which causes throttling and poor gaming performance.

Gigabyte HD 7850 1GB Windforce
Intel Core i5 3470
Gigabyte Z77
CX430

It all stated after I disassembled the graphics card (for the first time completely) to clean it. I removed the original thermal interface material and reapplied with another type, because that is the only one I could find to buy.

It is annoying since it is all perfectly clean but temperature rises to 102º C and then the GPU starts to throttle its performance due to the extremely high temperature. I decided to clean the whole computer because I thought the great amount of dust over it was influencing on overall components temperature.

I guess it is not possible to blame only the cheap TIM I applied since 102º C looks like even the heatsink cannot provide any cooling to the chip.

I have already reapplied the thermal paste in every way my mind could think of so far, and even when it is well applied and the heatsink pressure seems enough the temperature rises to the maximum in a few seconds.

I don't feel like buying online those expensive thermal compound because I am sure there is no way the original one can make this difference of 40º C.

What is wrong?

Thanks a lot for your help!

 

[Juliuz]

Anyone, please?

 

[Juliuz]

Help, please...

 

[pete_101]

You say the temperature rises to the max in only a few seconds, sounds like something is stopping the heat sink from making contact to the chip. When you remove the heat sink, does it look like the thermal paste has made even contact and has been pressed into a thin layer? If it has then you shouldn't be getting this problem. If not, carefully examine what may be stopping it from making a good contact.

One other thing, what drivers are you using? I know AMD were cooking cards a few months back when Crimson first came out.

 

[Juliuz]

Hey pete, thanks for answering. Furthermore, I apologize for my delay in replying.

So yeah, I have paid attention to that specially: after applying thermal paste and installing the heatsink, I always remove it again just to see how much pressure the compound has received between the surface. And it looks ok, since the heatsink has two copper heatpipes that seem a bit low compared to the aluminium surface. That gap between the chip and the copper heatpipes redirects the thermal paste and then is filled with considerable larger amount of it. Could it be the problem?

I am really out of ideas, since there's nothing I can do to prevent the gpu from decreasing it performance due to 102ºC temperature. And I realised it just happens when some application makes full usage of the gpu (Euro Truck Simulator 2 as example); until 70% (of gpu usage) the temperature stabilizes at 80ºC, but if past that, throttling happens in few seconds after reaching 100º C+.

Do you guys think the original (manufacturer) TIM had a key role in conducting heat to the copper heatpipes?

I will have to afford to buy a new card if I want to play gpu demanding games, until then I am stuck with Rocket League which is the only one who doesn't bring about the throttling issue (25FPS in any other game).

Thanks a lot for the help so far!