Lecture 3/12: Bits, Bytes and Words

Main ideas / concepts

Basic building blocks of computer memory

• Bit: stores exactly one value: 0 or 1.
• Bits are grouped into larger units:
  • Byte: the most common base unit for memory size
    • 1 byte = 8 bits
  • Word: typically (and recommended to assume)
    • 1 word = 4 bytes = 32 bits
  • Halfword / short
    • 1 halfword = 16 bits = 2 bytes
  • Double word
    • 2 words = 8 bytes = 64 bits
  • Quad word
    • 16 bytes

Terminology caution: “word” is not universal

• In most modern usage: word = 4 bytes / 32 bits
• In older microprocessors, word was sometimes 16 bits
• Lesson: to avoid confusion, explicitly say “word is 32 bits / 4 bytes.”

Binary vs hexadecimal: why humans prefer hex

• There is a 1-to-1 correspondence between:
  • hexadecimal digits and binary groups of 4 bits (nibbles)
• Using binary directly is error-prone for humans (too many 0s and 1s).
• Therefore, memory contents are usually expressed in hexadecimal.

Mapping hex digits to bytes/words (practical counting rule)

• Byte = 8 bits = 2 hex digits
• Word (32 bits) = 8 hex digits = 4 bytes
• Rule of thumb:
  • Count hex digits → determine how many bits/bytes are represented
  • 4 hex digits = 16 bits = 1 halfword
• Formatting guidance:
  • Hex digits are typically written in pairs (rarely anything smaller than a byte).
  • Often separate groups by spaces or dashes (commonly one byte per group).

Nibble

• A nibble is a 4-bit unit (half a byte).
• Mentioned as a “funny” term used less in serious contexts.

---

Method / instruction-style points

How to interpret a hex value’s size

• For a byte: 2 hexadecimal characters (2 hex digits)
• For a halfword (16-bit): 4 hexadecimal characters
• For a word (32-bit): 8 hexadecimal characters
• In general:
  • Each hex digit represents 4 bits
  • Each byte (8 bits) therefore takes 2 hex digits

How to format hex for readability

• Write hex digits in pairs (treat each pair as one byte)
• Use spaces or separators to keep byte boundaries visible
• If lots of digits:
  • Separate per byte with a dash every two characters

---

Octal number system (historical notes + pitfalls)

Base and correspondence

• Octal is base 8
• Digits range 0–7
• Each octal digit corresponds to 3 binary bits

Why octal is not used much

• Grouping problem with bytes
  • Octal groups are 3 bits, which does not align cleanly with 8-bit bytes
  • Some groups cross byte boundaries, making interpretation harder
• Numbers become longer
  • Because 3-bit grouping yields more digits than hex (4-bit grouping)

Why octal looks misleading

• Octal numbers can resemble decimal
  • Example: 263 (as octal) = 179 (decimal) (not 263)
• Leading zero confusion
  • In C, a leading 0 means the number is octal, not decimal
  • In math, leading zeros are often considered insignificant, but in C they change meaning
  • Adding a leading 0 can change the value from decimal 263 to octal 263

---

C integer sizes on Intel (and related guidance)

Architectures mentioned

• IA-32: 32-bit Intel architecture
• x86-64: 64-bit architecture

Primary integer types (workhorse and common sizes)

• int
  • Standard workhorse
  • Typically 32 bits = 4 bytes
• short
  • 16 bits = 2 bytes (halfword)
• char (or “chars” mentioned; auto-caption artifact)
  • 8 bits = 1 byte

Larger integer types

• long and long long
  • Lecturer’s recommendation:
    • Never use long (because on Intel compilers it may be 4 bytes or 8 bytes depending on environment)
    • long long usually means 8 bytes (i.e., a double word)

Floating point

• Mentioned as a separate topic (representation will be discussed later)

Pointers

• Pointer sizes depend on architecture:
  • 32-bit architecture: pointers use 4 bytes
  • 64-bit architecture: pointers use 8 bytes
• Example mentioned: char * meaning “pointer to characters (or generally to something)”

Type-name flexibility in C

• short can also be written as short int
  • Still the halfword type
• long can also be written as long int
  • Meaning depends on compiler/machine
• long long can sometimes be written as long int
  • Noted as less common

---

Speakers / sources featured

• No specific named speaker is identified in the subtitles.
• Source: A YouTube lecture titled “Lecture 3/12: Bits, Bytes and Words” (auto-generated subtitles).