Abstract

Floating point arithmetic is one of several methods of performing computations in digital designs; others include integer and fixed point computations. Fixed point utilizes a method comparable to scientific notation in the binary domain. In terms of computations, floating point is by far the most prevalent in today’s digital designs. Between the support offered by compilers, as well as for ready-to-use IP blocks, floating point units (FPU’s) are a de-facto standard for most processors. Despite its prevalence in modern designs, floating point has many flaws. One of the most common is the use of not-a-numbers (NaN’s). These are meant to provide a way of signaling invalid operation, however the excessive amount of them wastes usable bit patterns. As an alternative to floating point, a system named "Universal Numbers" or UNUMs was developed. This system consists of three different types, however for hardware compatibility, the Type III provides the best stand in for floating point. This system eliminates the NaN problem by only using one bit pattern, and also provides many other inherent benefits.

Publication Date

5-2020

Document Type

Master's Project

Student Type

Graduate

Degree Name

Electrical Engineering (MS)

Department, Program, or Center

Electrical Engineering (KGCOE)

Advisor

Mark A. Indovina

Advisor/Committee Member

Sohail A. Dianat

Campus

RIT – Main Campus

Download

Share

COinS