Originally Posted by wiredg
I'm confused. According to this nvidia page
: "NVIDIA GPUs are designed to operate reliably up to their maximum specified operating temperature. This maximum temperature varies by GPU, but is generally in the 105C range (refer to the nvidia.com product page for individual GPU specifications)."
OK, so let's go to the related nvidia product page. It claims that the GTX 680 max temp is 98C
So then, why does the company prefer to throttle this chip at 70C? Obviously, we want to avoid the max temp, but why shy away from it by such a wide margin? There ought to be some headroom between 70C and 98C. Why isn't 80C or 85C a perfectly safe limit during hard use like gaming?
Not 100% sure with this, but 105° is probably the thermal limit, going over this value has a high chance of damaging the gpu rapidly. Under 105° is considered safer since they are reasonably sure the gpu will not die in a matter of hours at that temp.
Although 105° is the max allowable temp, it isn't good for the gpu to be too hot, so they set a throttle point at 70° to help keep the card cooler, they are pretty sure that the majority of cards will live through their warranty period sticking at that value.
Between 70° & the 105° max won't kill the card, but it may be that nvidia's research has shown that a larger percentage of warranty claims will be made if the cards run at say 80° for extended periods.
Nvidia makes those calls based on the reference cards, the ln2 bios on the lightning not throttling was MSI's decision.
As far as overclocking goes, it has always been the cooler the better. 80°C may be called the safe temp for a gpu, that just means it won't die quickly, if it is considerably cooler it will overclock better.