nibblepacked

Nibblepacked is a term used to describe data that is stored by packing two 4-bit nibble values into a single 8-bit byte, creating a dense representation for small integers or hexadecimal digits. It is not a formal standard, but a descriptive concept found in documentation and codebases that require space-efficient storage.

In a nibblepacked sequence, each byte comprises a high nibble and a low nibble. For example, the byte 0x3A encodes the two nibbles 3 and A. Manipulation typically relies on bitwise operations: high = (b >> 4) & 0x0F; low = b & 0x0F. To pack two values v1 and v2, the byte is formed as ((v1 & 0x0F) << 4) | (v2 & 0x0F).

Applications of nibblepacking include compact storage for BCD digits, streams of hexadecimal digits, sensor data in embedded systems, and network or file formats where bandwidth or memory is limited. By storing two 4-bit units per byte, data density increases at the cost of extra processing to pack and unpack values.

Advantages of nibblepacked representations include reduced memory usage and potential improvements in cache efficiency. Limitations include added CPU overhead for packing and unpacking, possible readability and maintainability challenges, and the need for careful handling of value ranges and per-byte endianness considerations. In practice, nibblepacking is implemented wherever tight data density is prioritized and the extra bitwise logic is acceptable.

See also: nibble, bit packing, BCD, hex encoding.