https://www.sciencedaily.com/releases/2017/03/170328082917.htm http://news.stanford.edu/2015/12/09/n3xt-computing-structure-120915/ This is related to my earlier post about liquid computers. Like anything else on chips these days, they are now seriously investigating shrinking the cooling system as well with graphene or any number of different thermoelectric materials, possibly, even sandwiched between layers with up to 64 chips being stacked on top of one another today. Already graphene is used in lithium ion batteries for this purpose and there's no reason it can't also provide power and broad bandwidth communications to all the chips while it cools them. An interesting aspect of this which the article doesn't touch on is that chips are increasingly being designed to use less power in the first place making them naturally cooler. For example, AMD's Raven Ridge chips coming out later in the year should be capable of producing the equivalent power of a PS4 on laptops using a mere 35 watts and their next generation graphics architecture known as "Navi" may be capable of doing the same or better at around 10 watts. Theoretically, otherwise pretty ordinary electronic chips might be capable of going as low as about .03 watts which would produce a negligible amount of heat and many might assume it would be pointless to cool such chips, but nothing could be further from the truth. The smaller they become the more heat sensitive and the better the cooling system the more chips you can stack on top of one another with something like an exoscale computer fitting into something perhaps a bit larger than a Rubic's cube. If just every other layer of a 64 layer chip were to contain a cpu or gpu that means something the size of a contact lens could contain a computer roughly 30 times more powerful than any laptop today using already existing chips modified for the task. Of course, one of the alternatives being explored is using graphene nanotubes to do the processing as well with Stanford already having designed a "skyscraper" chip of carbon nanotubes that is a thousand fold more powerful than anything on the market today. However, there are a wide variety of alternatives being explored including 270 terahertz plasmonic chips and spintronics which are also making excellent progress as well. The advantage of graphene nanotubes is that research into them has been extensive in recent years and the technology to begin mass producing them already looks feasible at this point. What has been holding back this kind of technology is the willingness of anyone to invest billions in a processing plant without any assurances that some other technology won't come out and make it obsolete overnight, but the more progress they make towards producing a complete package, containing even its own cooling system, the more likely investors will jump on the bandwagon and kick-start the next technological revolution.
