Title :
The Quest for Reliable Nano Computations
Author_Institution :
Chair Computer System Engineering www.eecs.wsu.edu/~vbeiu Centers for Neural Inspired Nano Architectures www.cnina.org UAEU - College of IT, PO Box 17555, Al Ain, UAE www.cit2.uaeu.ac.ae Phone +971 (3) 713-3690 Fax +971 (3) 762-6309
Abstract :
In this presentation, we explore the feasibility of designing reliable nano-architectures using practical (i.e. very small = "less than 10") redundancy factors. To this end, we begin with a thorough review of redundant design strategies for fault-tolerant nano-architectures. We then adapt three redundant design strategies -- modular redundancy, von Neumann multiplexing, and reconfigurability -- to majority-gate circuits, and analytically evaluate these designs\´ reliabilities for very small redundancy factors (including fractional factors), using arguments as needed. This analysis motivates several extensions that allow optimization of reliability for very small redundancy factors, and highlights the benefit of using majority-gates in nano-scale design, paving the way for practical fault-tolerant architectures. Besides reliability, we simultaneously address low-power designing, and show that high-performance circuits can be operated reliably at ultra low switching energies.
Keywords :
Artificial neural networks; Biological neural networks; Circuits; Computer networks; Design optimization; Educational institutions; Fault tolerance; Redundancy; Societies; Systems engineering and theory;
Conference_Titel :
Microelectronics, 2005. ICM 2005. The 17th International Conference on
Print_ISBN :
0-7803-9262-0
DOI :
10.1109/ICM.2005.1590021