Zeroextending

Zero-extending, or zero extension, is the operation of increasing the width of a binary value by padding zeros in the most-significant bits, thereby preserving the value when the number is treated as unsigned. The original n-bit value is mapped to an m-bit value (m > n) by inserting (m − n) zeros at the high end. This is distinct from sign extension, which pads with the most-significant bit (the sign bit) to preserve the value in two's complement signed representations.

Example: extending an 8-bit value 0xCA (11001010 in binary) to 16 bits yields 0x00CA. The numeric value remains 202 if interpreted unsigned, even though the width has increased.

In hardware and software, zero-extension is common when widening unsigned integers or converting between different Word sizes. In CPU architectures, it is frequently implemented to prepare operands for wider arithmetic or to unify widths for bitwise operations. In x86 assembly, instructions such as MOVZX perform zero-extension, for instance moving an 8-bit register into a 32-bit destination with zeros filled in the upper bits.

In high-level languages, converting a smaller unsigned integer to a larger unsigned type typically results in zero-extension, while converting signed types may involve sign-extension depending on language rules. Zero-extension is essential for preserving unsigned magnitudes during type widening, memory addressing, and mixed-width computations. It cannot represent negative values; for negative numbers, sign-extension or other signed representations must be used.