Resource requirements for digital computations on electrooptical systems (original) (raw)

In this paper, we study the resource requirements of electro-optical organizations in solving computationally intensive problems such as 2-D image convolution. Using a generic model of parallel computation with optical interconnects called OMC, we derive the relationships for information transfer versus space/time tradeo s in solving a problem. Irrespective of the I/O scheme and the order of computation, we show that a lower bound of (nw) memory space represents the minimum hardware required for convolving a w w kernel with a n n image, if the input bits are given to the system only once.