GPU Throttles to 500 MHZ due to overheating but never recovers until I restart

penn919

Distinguished
Aug 24, 2010
293
5
18,865
I have a Radeon HD 7950 PowerColor PCS+ Edition that I bought on a forum a couple of years ago. The owner used the card for mining. I mainly play older games so I didn't notice any performance issue until one day while playing Tomb Raider I noticed the fps would suddenly dip down to an average fps of 27.

I soon figured out that the card has an overheating issue which I'm trying still trying to resolve (Applying new thermal paste on the GPU didn't work). My main question is with respects to the throttling behavior. The GPU would throttle down to 500MHZ whenever the temperature reaches 102 C. Obviously the temperature would dive. While under load the temp would be 71 C and while idling the GPU core would get as low as 40C.

The problem is that the normal core clock would never recover. The limit remains 500MHZ until I restart the computer. Is that normal behavior or is there someway to fix that?

Bonus question:
Would mining cause a GPU to have permanent overheating issues? The fans work and I applied new thermal paste. I even blew out all the dust on the heatsink and the fans. It made no difference whatsoever.

 

penn919

Distinguished
Aug 24, 2010
293
5
18,865
I decided to try removing my side panel and laying my mini tower on its side. I ran the same tomb raider bench mark maxed out for an hour. The GPU now only maxes out at 77 C very briefly then goes down to about 75 c. That's over 25 degrees difference! Is the side panel supposed to have that much of an affect?