New Nanoscale Material Could Allow Computer Chips To Rewire Themselves
The size of computer chips has been shrinking at a rapid pace over that last decade or so. While that gives us awesomely compact and capable devices like tablet PCs and smartphones, the issue is that we’re coming up on the the logical end of this advancement; soon we won’t be able to shrink things any smaller.
Scientists at Northwestern University have not found a way around this problem exactly, but they have been developing a nanoscale material that could allow chips to become more powerful and more efficient without necessarily making them smaller. Instead, this material would allow chips to effectively rewire themselves on the go in order to increase the efficiency of whatever process happens to be executing at the moment.
The material is made out of a bunch of electrically conductive, positively charged particles five nanometers wide surrounded by another bunch of negatively charged atoms, which can be rearranged for whatever purpose is needed. If that isn’t easy to picture, Bartosz A Grzybowski, a lead scientist on the project, described it this way
Like redirecting a river, streams of electrons can be steered in multiple directions through a block of the material – even multiple streams flowing in opposing directions at the same time.
Basically, what this means is that units of the material can be used as a diode, then changed to a resistor, then back to a diode, then maybe to a transistor. The process is both dynamic and reversible.
Now this technology is still in its infancy. The article outlining its existence was only published today. Still, should it gain some momentum — and I don’t see why it wouldn’t — it could help us to continue improving the efficency of computer chips even as we start to come up against the atomic-sized limit. It’s a good thing too, because unless we start coming up with some more clever ways to improve performance besides cramming things closer together, we’re going to run into a wall until someone can figure out quantum computing and that could be, well, hard.
(via The Register)