DIVISION UNIT WITH NEWTON-RAPHSON APPROXIMATION AND DIGIT-BY-DIGIT REFINEMENT OF THE QUOTIENT

The advantages of the convergence with the square of the Newton-Raphso n method are combined with the precision characteristics of digit-by-d igit algorithms to obtain units for fast division that satisfy the IEE E 754 floating point standard requirements. A general design methodolo gy that leads to a class of alternative architectures providing intere sting performances for division is presented, together with one exampl e of possible implementation. In particular, the proposed implementati on achieves a speedup varying from 20% to about 30% in comparison with a previous architecture by Fandrianto, with a relatively small additi onal hardware cost if a multiplier is already available on the arithme tic unit.