There’s been speculation recently that we’re nearing the end of moore’s law, which describes the exponential growth of computational power that we’ve seen over the past 30 or 40 years. There are a number of issues, chief among them being heat – the silicon wafers that CPU’s are created from are approaching the point where the density of transistors on them is generating heat close to their melting point. One approach to this is cooling – ever more sophisticated cooling systems to handle the heat burden, but as a practical matter it’s probably not feasible to get a liquid nitrogen cooling system installed in those p5 systems Dell would love to sell you in a few years.

As a result this is an area of intense research. What may surprise you is that the holy grail of semiconductor research, at least according to most folks, is diamonds. Diamonds can handle the most extreme heat. But of course they’re exorbitantly expensive, and I don’t think diamonds 1 foot or larger in diameter even exist, and that’s the size you would need to slice wafers from to build semiconductors on. So what’s a researcher to do? Make artificial diamonds of course! Industry has already been doing this for decades but not using technologies which are economical or capable of producing those 1 foot wafers you need to build general purpose semiconductors.

Until now. To most everyone’s surprise, not one but two companies appear to be close to solving this problem. Already one of them is beginning commercial shipments of artificial diamonds to the jewelry industry that are largely indistinguishable from mined diamonds, and their intent is to use the capital generated from these sales to fund semiconductor-class diamond production.

This is fascinating stuff. Mind you, none of this is likely to lead to a diamond-powered p5 sitting on your desk next year. But a diamond-powered p10 is now within the realm of the possible, wereas if you had asked scientists about this 6 months ago almost all of them would have said it’s basically science fiction. This is one of those huge huge scientific breakthroughs that basically flys under the radar. No one will notice this has happened and yet it is very likely to cause a fundamental shift in our underlying computing technologies over the next 20 years or so. Cool cool beans, and 2000 frames a second in Quake 5 😉

If the subject interests you, wired magazine has a great in-depth piece on it.