Blog: In Chaotic Computing, Anarchy Rules OK

In Chaotic Computing, Anarchy Rules OK
New Scientist (10/29/08) No. 2860, P. 40; Graham-Rowe, Duncan

Building next-generation computer processors by tapping the electronic parallel of chaotic weather systems is the goal of a team of physicists in the United States and India led by William Ditto of the University of Florida in Gainesville. Such processors would be vastly more powerful than their conventional chip equivalents, as well as self-reparable, through their ability to channel all their computational muscle into the task at hand and then reassign it as soon as a different chore comes up. The unpredictability of chaotic systems is the result of their sensitivity to the most infinitesimal influences, which inspired Ditto and Sudeshna Sinha of the Institute of Mathematical Sciences in Chennai to consider the construction of a circuit that exhibited chaotic behavior that could be harnessed for practical applications. Ditto and Sinha conceived of a chaotic logic gate with two inputs and one output like a conventional gate, but composed of a chaotic element or chaogate. When the chaogate receives its input signals, the internal chaotic circuit starts oscillating and quickly stabilizes at a value that relies on the inputs and a control signal. The research team calculated that changing the control signal's setting would enable the chaogate to be transformed into any desirable logic gate, and a prototype chaogate proved the concept's feasibility. Ditto is currently engaged in the commercialization of the technology and the fabrication of prototype circuits, and one of the promised advantages of chaotic logic is the dramatically reduced cost of producing custom chips. If a chip containing chaotic logic gates suffers damage, performance need not be affected as the circuits can be reconfigured to bypass the damaged area. Ditto's team has developed a method to use "chameleon" logic circuits to store data, producing digital memory that offers greater compactness than conventional memory and that also can retrieve data faster.

View Full Article