PLAM: a Posit Logarithm-Approximate Multiplier

TL;DR

This paper introduces PLAM, a novel approximate multiplier for Posit numbers that significantly reduces hardware complexity and power consumption in neural network applications without sacrificing accuracy.

Contribution

The paper presents a new logarithm-approximate multiplication scheme for Posit numbers, improving efficiency over existing multipliers in neural network hardware.

Findings

Reduces multiplier area by up to 72.86%

Decreases power consumption by up to 81.79%

Lowers delay by up to 17.01%

Abstract

The Posit Number System was introduced in 2017 as a replacement for floating-point numbers. Since then, the community has explored its application in Neural Network related tasks and produced some unit designs which are still far from being competitive with their floating-point counterparts. This paper proposes a Posit Logarithm-Approximate Multiplication (PLAM) scheme to significantly reduce the complexity of posit multipliers, the most power-hungry units within Deep Neural Network architectures. When comparing with state-of-the-art posit multipliers, experiments show that the proposed technique reduces the area, power, and delay of hardware multipliers up to 72.86%, 81.79%, and 17.01%, respectively, without accuracy degradation.