Abstract
The architecture of hypercubes is explained, and its advantages over other parallel architectures are examined. The flexibility of the interconnect scheme is discussed. The efficiency of hypercubes and hardware aspects are considered.