Based on the Newton method, a table driven iterative scheme is proposed to compute the logarithm function. The proposed scheme provides high performance and high precision and it is implementable with currently available technologies. It is shown, assuming the availability of a parallel multiplier and an adder and the IEEE floating point format, that a computation delay of 12 machine cycles can be obtained with precision of 0.761×2-53 requiring 18432 bytes of read only memory (ROM). Based on performance and look-up table size estimations of other existing schemes also using table driven approaches a speed-up of over 1.5 for the proposed scheme can be obtained
