A binary algorithm for solving n by n systems of integer equations (Q2266352)

We discuss a ''binary'' algorithm for solving systems of linear equations with integer coefficients. So-called ''binary'' algorithms differ from ordinary ones in that there is no roundoff error, but only overflow, and the underlying analysis is p-adic analysis rather than conventional real analysis. The advantages of this algorithm are especially apparent when extremely large numbers are involved and roundoff error can be tolerated. VLSI implementation of this and other ''binary'' algorithms is very appealing because of the extreme regularity of the circuits involved.