The semiconductor industry is consistently looking at miniaturisation, reduced power, and enhanced levels of sophistication driven by the needs of application domains such as consumer electronics.
Advances such as strained silicon and silicon-on-insulator technologies are just two examples that illustrate the industry’s interest in exploring new dimensions of growth that differ from well-established platform solutions.
A research group at Rice University in the US has now looked beyond boolean logic for electronic devices and has used probabilistic logic to develop a microchip that is said to consume about 30 times less power while running seven times faster than existing solutions. The technology behind this has been dubbed probabilistic complementary metal oxide semiconductor (PCMOS) technology.
PCMOS was jointly validated by Rice and Nanyang Technological University (NTU) in Singapore via a joint institute founded in 2007, the Institute for Sustainable Nanoelectronics (ISNE).
Technologists associated with the research cite that, beyond proving itself on the power and speed metrics, the solution is in line with the expectation of physicists in the technology space who feel that the next generation of processors will increasingly rely on probabilistic logic. The new probabilistic approach to signal processing has resolved the fundamental noise challenge which arises when developers strive to miniaturise their chips.
As chips get smaller, the impact of noise gets stronger. This eventually led to enhancing the operating voltage in a direction to overpower noise, which again restrains the ability to reduce power. For many years, engineers have been striving to break this imbalance and now PCMOS’s unique approach has potentially, efficiently resolved this challenge. The platform lowers the voltage dramatically and deals with the resulting computational errors by embracing the errors and uncertainties through probabilistic logic.
The prototypes were application-specific integrated circuits, or ASICs, that were designed solely for encryption. The Rice-NTU team plans to follow its proof-of-concept work on encryption with proof-of-concept tests on microchips for cellphones, graphics cards and medical implants.
Accepting that it is impossible to eliminate errors in systems has been very useful for addressing key challenges in the industry. Domains such as digital signal processing (DSP) have evolved to what they are today only because researchers explored ways to use or work around errors in a system by incorporating suitable approximation algorithms, rather than trying to eliminate them.
The success of these DSP solutions could be considered a key marker that promises success for solutions such as PCMOS. Further, it is expected that, since the whole platform is built around CMOS, establishing a manufacturing platform for PCMOS devices should be relatively easy.
With reduced power consumption being one of the dominant promises of PCMOS, one could expect this technology to contribute substantially toward going green, with the first application in embedded computing possible in a period of four years.
For more information contact Patrick Cairns, Frost & Sullivan, +27 (0)21 680 3274, [email protected],
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version