IBM is putting $3 billion into two projects designed to rethink the very concept of computer chip production. It says such a move is needed to meet increasing demands for processing without impractical increases in energy use.
Both projects will run for five years. They both involve exploring the continuing efforts to shrink chips — or more specifically, to allow more processing power from a single chip.
Chips are usually referred to by their half-pitch measurement, which is half the distance between identical points. (In simple terms the smaller the half-pitch, the denser the circuitry on the chip and the greater the processing power.)
As a general rule, a new generation of chips comes along every couple of years with a half-pitch around 40 percent smaller than its predecessor. That’s led to drops from 800 nanometers in the late 1980s to 22 nanometers being common today. The current schedule would hit 7 nanometers by the end of this decade.
The problem is that we’re rapidly hitting the point where more efficient manufacturing may not be enough to keep up the pattern. The current technique of etching purified silicon may not be suitable for ever-smaller circuitry, which is why IBM wants to explore alternatives.
The first of the two projects will look at ways of solving the problems with smaller silicon circuits, while the second will look at whether a completely different chip manufacturing approach might be better.
Network World notes this could involve ditching silicon and using cylinders of carbon atoms, known as carbon nanotubes, though this has both practical and safety questions.
Other topics IBM will explore include:
- quantum computing (which could be described as a 3D version of binary);
- neurosynaptic computing (mirroring the structure of the brain with its highly efficient processing power);
- silicon photonics (trying to recreate the benefits of fiber-optic over copper cabling but on a tiny scale; and
- graphene (because whatever the question in science and technology, graphene is always on the list of possible answers.)