What's the difference between a word and byte?

Question

I\'ve done some research. A byte is 8 bits and a word is the smallest unit that can be addressed on memory. The exact length of a word varies. What I don\'t understand is wh

Answer 1

The exact length of a word varies. What I don't understand is what's the point of having a byte? Why not say 8 bits?

Even though the length of a word varies, on all modern machines and even all older architectures that I'm familiar with, the word size is still a multiple of the byte size. So there is no particular downside to using "byte" over "8 bits" in relation to the variable word size.

Beyond that, here are some reasons to use byte (or octet¹) over "8 bits":

1. Larger units are just convenient to avoid very large or very small numbers: you might as well ask "why say 3 nanoseconds when you could say 0.000000003 seconds" or "why say 1 kilogram when you could say 1,000 grams", etc.
2. Beyond the convenience, the unit of a byte is somehow as fundamental as 1 bit since many operations typically work not at the byte level, but at the byte level: addressing memory, allocating dynamic storage, reading from a file or socket, etc.
3. Even if you were to adopt "8 bit" as a type of unit, so you could say "two 8-bits" instead of "two bytes", it would be often be very confusing to have your new unit start with a number. For example, if someone said "one-hundred 8-bits" it could easily be interpreted as 108 bits, rather than 100 bits.

---

¹ Although I'll consider a byte to be 8 bits for this answer, this isn't universally true: on older machines a byte may have a different size (such as 6 bits. Octet always means 8 bits, regardless of the machine (so this term is often used in defining network protocols). In modern usage, byte is overwhelmingly used as synonymous with 8 bits.

Answer 2

A group of 8 bits is called a byte ( with the exception where it is not :) for certain architectures )

A word is a fixed sized group of bits that are handled as a unit by the instruction set and/or hardware of the processor. That means the size of a general purpose register ( which is generally more than a byte ) is a word

In the C, a word is most often called an integer => int

Answer 3

The terms of BYTE and WORD are relative to the size of the processor that is being referred to. The most common processors are/were 8 bit, 16 bit, 32 bit or 64 bit. These are the WORD lengths of the processor. Actually half of a WORD is a BYTE, whatever the numerical length is. Ready for this, half of a BYTE is a NIBBLE.

Answer 4

It seems all the answers assume high level languages and mainly C/C++.

But the question is tagged "assembly" and in all assemblers I know (for 8bit, 16bit, 32bit and 64bit CPUs), the definitions are much more clear:

byte  = 8 bits 
word  = 2 bytes
dword = 4 bytes = 2Words (dword means "double word")
qword = 8 bytes = 2Dwords = 4Words ("quadruple word")

Answer 5

Whatever the terminology present in datasheets and compilers, a 'Byte' is eight bits. Let's not try to confuse enquirers and generalities with the more obscure exceptions, particularly as the word 'Byte' comes from the expression "By Eight". I've worked in the semiconductor/electronics industry for over thirty years and not once known 'Byte' used to express anything more than eight bits.

Answer 6

In this context, a word is the unit that a machine uses when working with memory. For example, on a 32 bit machine, the word is 32 bits long and on a 64 bit is 64 bits long. The word size determines the address space.

In programming (C/C++), the word is typically represented by the int_ptr type, which has the same length as a pointer, this way abstracting these details.

Some APIs might confuse you though, such as Win32 API, because it has types such as WORD (16 bits) and DWORD (32 bits). The reason is that the API was initially targeting 16 bit machines, then was ported to 32 bit machines, then to 64 bit machines. To store a pointer, you can use INT_PTR. More details here and here.