Principal Investigator Neil Gershenfeld
Co-investigator Erik Demaine
Project Website http://rala.cba.mit.edu/
RALA is a new approach to computer architecture.
Computer architecture hasn't changed substantially since 1945, but the ways computers are manufactured and used have changed completely. As a result, computer clocks have not gotten faster since 2004. The problem is that the current abstract model of computation hides too much of the physical reality from the software world, placing extra burdens on the hardware designers as computing systems get more and more powerful. Since RALA acknowledges physical causality, the burden on the hardware designer remains constant independent of the total system size. RALA is a scalable mathematical model that can be applied to a diverse class of physical technologies, from ordinary silicon chips to newer technologies such as memristors, crossbar latches, single-electron transistors, and others, to aid the next paradigm shift in computing technology.