Reconfigurable cube architecture for parallel computation

The construction of a dynamic cube network architecture that extends the capabilities of hypercube architectures with only a small increase in hardware complexity is described. The network is self-routing in the sense that there is a simple distributed routing algorithm which guarantees optimal paths between any pair of the vertices. This, together with other improvements in embedding properties, makes the dynamic cube an attractive alternative to the ordinary hypercube for massively parallel architectures.<
>