Researchers at Purdue University develop a special cooling system that extracts heat from all layers of the processor and brings it to the area of contact with the heatsink, to improve cooling.
Chips, to a greater or lesser extent, require cooling systems, especially those that perform intensive work and the better the cooling system, the longer the life of the processor itself. A small-scale system has long been sought to improve their cooling or to be able to eliminate dedicated cooling systems. Purdue University, Lafayette, Indianapolis, would have been able to develop intrachip cooling technology, something that holds great promise for the processors and semiconductor industry in general.
This research has been developed by a DARPA team at the Birck Nanotechnology Center at the aforementioned University. This project started from a basic idea set by DARPA, who wanted to develop a technology that had the capacity to cool up to 1kW of heat per square centimeter, something that is ten times more heat than the heat sinks used for computer processors, like the one we have next to us.
Said cooling system has a cooling liquid, which is electrically isolated, which circulates inside the chips themselves, transported by microscopic channels. This system makes useless the IHS or integrated heatsinks, which simply extract the heat from the chip, but are unable to dissipate this heat. If we did not install a heatsink on the IHS, the processor would not start or it would burn almost instantly, because the heat would be trapped at this point.
"Today you can pack a lot of computing power into a very small chip," began Justin A. Weibel, professor of mechanical engineering at Purdue University, "So stack chips on top of each other in a 3D structure. it's the future. But of course, this presents a problem when it comes to cooling the chip, since the more layers there are, the more heat is generated. Normally you put a heatsink in the upper layer, but this leaves the lower layers with almost no cooling, so they are forced to reduce their performance to preserve their integrity ”.
Explained in a simple way, current processors make use of 3D structures and what is exposed is that when we put the heatsink, we only cool what is in the most superficial part of it, leaving the intermediate and lower zone without any type of cooling real. Through these channels the heat from the lowest layer is made to rise to the dissipation zone and through a heatsink we cool all the layers at the same time, not just one. This could allow to expand the number of processor layers and improve cooling, as well as the overclocking capacity of the processors.
Source: Purdue
