Infinity Representation in IEEE 754

IEEE 754 uses specific bit patterns to represent positive and negative infinity. These special values arise from operations that overflow or divide by zero, and they follow well-defined arithmetic rules.

Bit patterns for infinity:

Value	Float32 Hex	Float64 Hex	Binary Pattern
+Infinity	0x7F800000	0x7FF0000000000000	sign=0, exponent=all 1s, mantissa=all 0s
-Infinity	0xFF800000	0xFFF0000000000000	sign=1, exponent=all 1s, mantissa=all 0s

The key identifier: all exponent bits are 1, and all mantissa bits are 0. If even one mantissa bit were set, the value would be NaN instead.

How infinity arises:

• Division by zero: 1.0 / 0.0 = +Infinity, -1.0 / 0.0 = -Infinity
• Overflow: when a result exceeds the largest representable finite value (3.40e+38 for float32, 1.80e+308 for float64)
• Explicit assignment: Infinity, -Infinity, float('inf') in Python

Note that 0.0 / 0.0 produces NaN, not infinity.

Arithmetic with infinity:

Operation	Result
x + Infinity	Infinity (for finite x)
Infinity + Infinity	Infinity
Infinity - Infinity	NaN
Infinity * 0	NaN
Infinity * Infinity	Infinity
x / Infinity	0 (for finite x)
Infinity / Infinity	NaN

Comparison behavior:

• Infinity > x is true for any finite x
• -Infinity < x is true for any finite x
• Infinity === Infinity is true
• isFinite(Infinity) is false

Why infinity exists:

Instead of crashing on overflow, IEEE 754 returns infinity and allows computation to continue. This is often preferable to throwing an exception because downstream code can check for infinity and handle it gracefully. In signal processing and control systems, infinity can represent saturation naturally.