Gibibyte to Byte

GB (gigabyte) = 10⁹ bytes = 1,000,000,000 bytes (SI decimal). GiB (gibibyte) = 2³⁰ bytes = 1,073,741,824 bytes (IEC binary). GiB is 7.37% larger. This is why a 1 TB hard drive labelled by the manufacturer (using 10¹² bytes) appears as approximately 931 GiB in Windows or Linux (which divide by 1,073,741,824). Neither value is wrong; they use different counting systems.

Early PC games (1990s) fit on a few floppy disks — under 10 MiB. CD-era games (late 1990s) reached 650 MiB. DVD-era titles hit 4–8 GiB. Modern AAA games like Call of Duty or Flight Simulator now exceed 100–200 GiB due to uncompressed 4K textures, high-fidelity audio in multiple languages, and pre-rendered cinematics. The growth rate has outpaced Moore's Law: storage needs roughly double every 2–3 years for top-tier games, driven primarily by texture resolution increases that scale quadratically with pixel count.

A module sold as "16 GB" RAM by manufacturers means 16 × 10⁹ = 16,000,000,000 bytes? No — RAM is actually built in binary powers. A "16 GB" RAM module contains exactly 2³⁴ = 17,179,869,184 bytes = 16 GiB. In this case, the manufacturer is using "GB" to mean GiB — unlike hard drives, where manufacturers genuinely use decimal GB. RAM capacities are always powers of 2 in gibibytes.

A 512 GB SSD (decimal, as labelled by the manufacturer) holds 512,000,000,000 bytes. Divide by 1,073,741,824 to get GiB: 512,000,000,000 ÷ 1,073,741,824 ≈ 476.8 GiB. After OS overhead and firmware reserved space, the usable capacity shown in the OS is typically 450–465 GiB for a nominally 512 GB drive.

Yes — GiB is the technically correct unit for binary memory. RAM, CPU cache, and GPU memory are all physically organized in powers of 2, making GiB the natural unit. The JEDEC memory standard (the body that defines RAM specifications) officially uses the IEC GiB notation, even though product packaging often says "GB" for commercial reasons. In engineering and OS development contexts, GiB is the preferred term.

A byte contains exactly 8 bits. This is the universal modern standard, though early computing used variable byte sizes (5, 6, or 7 bits). The 8-bit byte became universal with the IBM System/360 in 1964. Eight bits allow 256 possible values (0–255), sufficient to encode all ASCII characters with room for control codes.

Eight bits became standard because it is the smallest power of two that can encode all 128 ASCII characters (7 bits) with a spare bit for parity checking or extended character sets. It also maps cleanly to two hexadecimal digits (0x00–0xFF), making it convenient for low-level programming and hardware design. Earlier systems used 6-bit or 7-bit bytes; 8-bit won due to IBM's dominance in the 1960s–70s.

A nibble (also spelled nybble) is 4 bits — half a byte. A nibble represents exactly one hexadecimal digit (0–F). The term is used in low-level programming, embedded systems, and BCD (binary-coded decimal) encoding. It is not an SI unit and rarely appears in general computing contexts outside of hardware and systems programming.

It depends on the character and encoding. In UTF-8 (the dominant web encoding): ASCII characters (A–Z, 0–9) use 1 byte; common European accented characters use 2 bytes; most Asian scripts (Chinese, Japanese, Korean) use 3 bytes; emoji and rare characters use 4 bytes. A plain English text file is efficiently encoded as 1 byte per character in UTF-8.

In most modern usage, byte and octet are synonymous — both mean 8 bits. "Octet" is preferred in networking standards (RFC documents, ITU specifications) to avoid ambiguity from early computing where byte sizes varied. Internet protocol headers are specified in octets; operating systems and storage devices use bytes. In practice you will encounter "octet" mainly in formal networking documentation.