A Two-Stage Operand Trimming Approximate Logarithmic Multiplier

We present an approximate logarithmic multiplier with two-stage operand trimming, which prioritises area and energy consumption while retains acceptable accuracy. The trims the least significant parts of input operands in first stage mantissas obtained operands' approximations second stage. evaluated multiplier's efficiency terms error, energy, utilisation using NanGate 45nm. experimental results show that proposed exhibits smaller than state-of-the-art designs it behaves well image processing classification convolutional neural networks.