As processing power increases, CPU transistor count goes up while the die size shrinks. Each new transistor require more energy and can limit the size of any one processor in a given system. Draining or charging the interconnecting wires between transistors adds transmission delay to the chip. Add this to the time it takes to flip and chain transistors states, a substantial time limit is places on the chip. These problems have plagued processors since the beginning.
An engineering team from Northwestern University state they have found a way around this problem through the use of ‘spin-logic circuits’. Unlike common circuits that house the transistors on silicon-based semiconductors, spin-logic circuits take advantage of a spin-polarized current sandwiched between a ferromagnetic substrate. This affects the spin of electrons which can then be manipulated into a logic circuit capable of using the Boolean-logic rules of traditional CMOS circuits with far less components in manufacturing and requires considerably less power requirements. As a result, the team was able to create their spin-logic circuit using magnetoresistive bipolar spin-transistors along with a magnetic resistive shielding (they do not specify as to what, but I’m guessing a Faraday/Kerr cage) which gives the transistors a cascade effect resulting in sophisticated function units. This means that, in theory, future CPUs using the team’s spin-logic circuits could be considerably faster and 1-million times more power efficient over today's CPUs. While their new circuit design is amazing, to say the least, incorporating them into new electronic devices such as processors or SSDs could take the better part of a decade or more.
Cabe
