G.M. Megson
X. Chen
University of Newcastle upon Tyne. 1993
This paper presents a method to map a computational polytope onto a given regular processor array. Beginning with a suitable space mapping matrix, the method constructs an activity matrix, proposed by Darte, according to the shapes of the computational polytope and the processor array. Then a valid timing vector is derived from a time basis which defines the activity matrix with the space mapping matrix together. By this method the given-shape mapping can be achieved without any difficulty. In addition, by designing the space mapping matrix and activity matrix meticulously, the method is easily implemented without exhausted computations.