Course contentsShow
← Computer Science
Lesson 1064 of 2,87225. Computer Architecture FoundationsFree lesson

Floating-Point Representation Overview

The concept of representing real numbers using separate fields for sign, exponent, and significand.

Floating-Point Representation Overview

You'll learn how computers represent real numbers by splitting them into three parts: sign, exponent, and significand.

The Core Idea

While integers are straightforward to represent in binary, real numbers (like 3.14159, 0.000002, or -1234567890.123) pose a challenge. Computers can't store infinite precision, so they use floating-point representation—a clever scheme that works like scientific notation.

Think of how we write very large or small numbers in science: 6.022 × 10²³ (Avogadro's number) or 9.109 × 10⁻³¹ (electron mass). Instead of writing out all those zeros, we separate the number into:

  • A sign (positive or negative)
  • A significand (the meaningful digits: 6.022 or 9.109)
  • An exponent (the power of 10: 23 or -31)

How It Maps to Binary

Floating-point numbers in computers work the same way, but in binary:

  • Sign bit: 1 bit indicating positive (0) or negative (1)
  • Exponent field: Several bits storing the power of 2 (not 10)
  • Significand (or mantissa) field: The actual significant digits in binary

For example, the decimal number -12.5 becomes:

  • Sign: negative (1)
  • Significand: 1.5625 in binary form
  • Exponent: 3 (because 1.5625 × 2³ = 12.5)

These three fields are packed together into a fixed number of bits, allowing computers to represent an enormous range of values—from astronomically large to infinitesimally small—with finite hardware.

Key Takeaway: Floating-point representation uses three separate fields (sign, exponent, and significand) to efficiently encode real numbers in binary, similar to how scientific notation works in decimal.