Abstract
Optical interconnection between VLSI circuits, computing chips or boards plays an essential role on parallel computing. Such an approach utilizes the fast switching speed of electronics and the massive parallelism and wide bandwidth of optics for communication and provides the potential of achieving extremely high speed [1,2,3]. A generalized optical crossbar switch [4] which can provide a reconfigurable and arbitrary interconnection between an array of N lasers and an array of N detectors can be implemented using a parallel matrix-vector multiplication scheme [5]. The N2 parallelism of this scheme is accompanied with an intrinsic fan-out energy loss [6]. The fan-out leads to an energy efficiency of 1/N.
© 1990 Optical Society of America
PDF ArticleMore Like This
Arthur E. Chiou, Pochi Yeh, and John Hong
FX1 OSA Annual Meeting (FIO) 1988
Arthur Chiou and Pochi Yeh
TuB2 Optical Computing (IP) 1989
Arthur E. Chiou and Pochi Yeh
TuD4 Optical Computing (IP) 1991