Title :
Ultra-Low power neuromorphic computing with spin-torque devices
Author :
Sharad, Mrigank ; Deliang Fan ; Yogendra, K. ; Roy, Kaushik
Author_Institution :
Dept. of Electr. & Comput. Eng., Purdue Univ., West Lafayette, IN, USA
Abstract :
Emerging spin transfer torque (ST) devices such as lateral spin valves and domain wall magnets may lead to ultra-low-voltage, current-mode, spin-torque switches that can offer attractive computing capabilities, beyond digital switches. This paper reviews our work on ST-based non-Boolean data-processing applications, like neural-networks, which involve analog processing. Integration of such spin-torque devices with charge-based devices like CMOS can lead to highly energy-efficient information processing hardware for applicatons like pattern-matching, neuromorphic-computing, image-processing and data-conversion. Simulation results for analog image processing and associative computing has shown the possibility of ~100X improvement in energy efficiency as compared to a 15nm CMOS ASIC.
Keywords :
CMOS analogue integrated circuits; electronic engineering computing; low-power electronics; magnetic domain walls; neural nets; spin valves; CMOS; ST-based nonBoolean data-processing applications; analog image processing; associative computing; current-mode switches; data conversion; domain wall magnets; information processing hardware; lateral spin valves; neural networks; pattern matching; spin transfer torque devices; ultra-low power neuromorphic computing; ultralow-voltage switches; Biological neural networks; Computer architecture; Magnetic domain walls; Magnetic domains; Magnetic tunneling; Neuromorphics; Neurons; analog; interconnect; logic; low power; neural networks; non-Boolean; programmable logic array; spin; threshold logic;
Conference_Titel :
Energy Efficient Electronic Systems (E3S), 2013 Third Berkeley Symposium on
Conference_Location :
Berkeley, CA
DOI :
10.1109/E3S.2013.6705865
