Parallel, iterative solution of large, sparse linear systems on hypercubes

A parallel, iterative algorithm for solving large, dense linear systems on hypercubes is presented. This algorithm is shown to be optimal, i.e. it results in minimum execution time for an iteration. The performance of the parallel algorithm is evaluated analytically and the results show that the hypercubes are highly effective in solving large systems of dense linear algebraic equations. Finally, the suitability of the hypercubes for solving sparse linear systems is discussed.<
>