This paper presents an e cient hardware algorithm for variable-precision logarithm. The algorithm uses an iterative te chnique that employs table lookups and polynomial approximations. Compared to similar algorithms, it reduces the number of xed-precisionoperations by avoiding full precision computations and dynamically varying the precision of intermediate results. It also uses signi cantly smaller tables than related algorithms. For a speci ed hardware implementation, the algorithm requires fewer than2L2 xe d-precision multiplications to evaluate the logarithm to L words of precision. An error analysis for the algorithm is also presente d.
Javier Hormigo, Julio Villalba, Michael J. Schulte