ABSTRACT
In scientific technology and engineering software, floating-point computing multiplication is a basic process. However, maintaining hardware resource efficiency while achieving high precision is a persistent challenge. This abstract presents a novel approach to the construction of a single-precision Inexact Floating-Point Multiplier. Our objective is to take into account the unique characteristics of the sign bit, exponent bits and mantissa bits are parts of single-precision floating-point numbers while striking a balance between computing accuracy and hardware efficiency. In this lecture, we will look at the design process, accuracy analysis, and potential applications of this imprecise multiplier, demonstrating its ability to perform workloads involving single-precision floating-point multiplication with noticeable efficiency advantages. This discovery opens up new applications for computing, including machine learning, digital signal processing, and scientific simulations.
