Nibble to Tebibit

A nibble is 4 bits, or half a byte. It encodes one hexadecimal digit (values 0–15, represented as 0–9 and A–F). Nibbles are important in BCD (binary-coded decimal) encoding, where decimal digits are packed two per byte (each digit occupying one nibble). Packed BCD is used in financial systems and legacy databases to represent decimal numbers without floating-point rounding errors.

Hexadecimal (base 16) maps perfectly to nibbles because 4 bits can represent exactly 16 values (2⁴ = 16). One byte = two nibbles = two hex digits. A byte value of 0xFF (255 in decimal) is two nibbles: F (1111) and F (1111). This mapping makes hexadecimal the natural notation for expressing binary data — programrs use hex because one hex digit always represents a fixed number of bits.

Binary-Coded Decimal (BCD) encodes each decimal digit (0–9) as a 4-bit binary value (nibble). Two decimal digits fit in one byte using "packed BCD". For example, the decimal number 47 is stored as 0100 0111 in packed BCD — each nibble holds one digit. BCD avoids the rounding errors of binary floating-point, which is why it is used in financial software, calculators, and legacy banking systems.

A nibble = 4 bits (1 hex digit). A byte = 8 bits (2 hex digits, 2 nibbles). A word = typically 16, 32, or 64 bits depending on the processor architecture (see the "word" unit for details). These are the fundamental granularities of digital data: nibble for hex/BCD, byte for text and addressing, word for native processor arithmetic.

Nibbles are rarely referenced directly in modern high-level programming but remain fundamental at the hardware level. Embedded systems, FPGA design, network packet parsing, and hardware description languages (VHDL, Verilog) regularly manipulate nibbles. The nibble is also the key concept behind hexdump utilities — the canonical way to inspect raw binary files and network packets.

A terabit (Tbit) = 10¹² bits (SI decimal). A tebibit (Tibit) = 2⁴⁰ bits = 1,099,511,627,776 bits (IEC binary). Tebibit is 9.95% larger. At enterprise storage scale, this 10% difference has real financial consequences: a storage specification error confusing Tbit with Tibit on a 100-unit deployment results in nearly 10 units' worth of capacity discrepancy.

Tebibits appear in: NAND flash memory die specifications and yield calculations, high-speed fabric interconnect specifications (InfiniBand HDR = 200 Gbit/s), supercomputer storage system designs, and academic papers on distributed storage systems. Consumer applications never display tebibits; the term is confined to engineering and procurement contexts.

Modern 3D NAND stacks 100+ layers of memory cells vertically. A single die from a 232-layer TLC NAND chip can hold about 1 Tibit (128 GiB) raw capacity. Manufacturers measure at the die level in tebibits because binary addressing maps directly to the physical array geometry — each layer, block, and page aligns to powers of 2. A 16-die package thus holds 16 Tibit (2 TiB) before error correction overhead.

Each binary prefix multiplies by 1,024 instead of 1,000. The compounding effect: kibi vs kilo = 2.4% difference, mebi vs mega = 4.9%, gibi vs giga = 7.4%, tebi vs tera = 9.95%, pebi vs peta = 12.6%, exbi vs exa = 15.3%. The difference grows by approximately 2.4% with each prefix step, making precision in naming increasingly important at larger scales.

1 Tibit = 2⁴⁰ bits = 2⁴⁰ / 8 bytes = 2³⁷ bytes = 137,438,953,472 bytes ≈ 137.4 GB (decimal). To convert Tibit to GB: multiply by 137.4. To convert Tibit to GiB: divide by 8 (since 1 Tibit = 0.125 TiB = 128 GiB). The exact value: 1 Tibit = 128 GiB.