ROM based logic design for base-2 exponential and logarithm converter using fixed point number representation

Abstract

Logarithm and exponential functions are frequently used in signal processing, communication and information theory. They are primarily used for hardware calculations, handling multiplications, divisions, powers, and roots effectively. This paper presents a ROM based approach for the evaluation of base-2 logarithm and exponential at single precision floating point input. Since the method used is non-iterative, the output of the logic has fixed accuracy. The accuracy has been improved as compared to the existing published non-iterative methods available for computing functions such as logarithms and exponential. The hardware is optimized for fixed point number representation using the suitable logic design. A combined logic to evaluate the logarithm and exponential for sustained input, which is implemented as single cycle and pipelined approach, is also explained. Finally, the error plots, hardware and delay reports are compared for single cycle and pipelined approaches.