A parallel implementation of the self-organizing feature map using synchronous communication

Parallel algorithms for implementing the Kohonen self-organizing feature map (SOFM) are presented. The algorithms can maintain a high degree of load balancing and minimize the communication overhead. The algorithms were implemented on a linear chain and a two-dimensional mesh of transputers. Significant speedup was achieved. Models to describe the performance of the algorithms are presented. The performance of massively parallel processing systems is predicted from the models.<
>