| 摘要 |
<p>The invention relates to a method and a circuit for processing digital signals representing vectors. The invention permits processing between pairs of vectors of vector sets (R, S) in order to ensure a spatial and temporal meeting of these vectors. This meeting is effected by parallel processing along a two-dimensional network (RE) of processing cells (CT) for the components attached to one dimension (x, y) of the network, and sequentially by partition of sets into sub-sets (Rv, Su) subjected to processing during the course of a microcycle, during which a sub-set (Su + 1) is memorised. The sub-set (Rv + 1) is memorised during a macrocycle with a view to the following processing and so on. Application to the product of vectors, vector matrix, and operations on databases. <IMAGE></p> |
