A memory is a long row of numbered boxes. Each box is a location, and the bits inside it are a word.
Every word in a given memory has the same width, the word size. A ×8 memory holds 8-bit words, a ×16 memory 16-bit words. The word is the unit of transfer: one read returns a whole word and one write replaces a whole word.
Keep three ideas apart:
- The address says which location. It is fixed by where the location sits.
- The word (the contents) is what is stored there. It changes whenever you write.
- The word size is how many bits each location holds. It equals the number of data lines.
Two locations can hold the same word; contents don't have to be unique, addresses do.
The number of words is a separate number again. In a 1K × 8 memory, 1K is the number of words (which fixes the address bits), and 8 is the word size.
Careful: "word" also means a CPU's natural data size (often 32 or 64 bits). In a memory question, a word is simply whatever one address holds, even a single bit or a byte.