OPTICAL MULTIPLEX COMPUTING BASED ON SET-VALUED LOGIC AND ITS APPLICATION TO PARALLEL SORTING NETWORKS

A new computer architecture using multiwavelength optoelectronic integ rated circuits (OEICs) is proposed to attack the problems caused by in terconnection complexity. Multiwavelength-OEIC architectures, where va rious wavelengths are employed as information carriers, provide the wa velength as an extra dimension of freedom for parallel processing, so that we can perform several independent computations in parallel in a single optical module using the wavelength space. This ''multiplex com puting'' enables us to reduce the wiring area required by a network an d improve their complexity. In this paper, we discuss the efficient mu ltiplexing of Batcher's bitonic sorting networks, highly parallel comp uting architectures that require global interconnections inherently. A systematic multiplexing of interconnection topology is presented usin g a binary representation of the connectivities of interconnection pat hs. It is shown that the wiring area can be reduced by a factor of l/r 2 using r kinds of wavelength components.