Computer words differ quite a bit from the regular meaning of word; it simply means four bytes of data. Historically this was the maximum size of numbers that could be added or subtracted in one step on a computer system, in a part of the electronics called a register.
Some computer systems have a different word length than four bytes, but these systems are rather unusual or maybe even outdated.
There is also a relationship between the register size and the lenght of an instruction, the binary codes that the processor actually executes. In many types of computers a good number of the instructions only need half the space as the others, hence the term half-word.
But back to the regular four byte variety of a word. Four bytes cannot hold a very big number--the biggest one possible is just 4,294,967,296, a little over 4 billion. If 4 billion seems like a lot, then consider that in 1997 the US budget was $1,579 BILLION or $1,579,000,000,000 which certainly will not fit in a computer word.
Multiplication is a bigger problem than addition because the size of the numbers can jump very quickly. 4,000,000,000,000 + 2 is just 4,000,000,000,002 and will still fit inside a word, but 4,000,000,000,000 x 2 is 8,589,934,592 and doesn't fit a word. To get around this problem a double-word using two adjacent registers is used. A double-word has can hold up to 18,446,744,073,709,551,616. Despite the enormous capacity, it won't be big enough for long enough as memory and storage ranges of computer systems jump.
The appearance of Quad-word or 128-bit use suggests that the common use of word and multiples will fade in favor N-bit processor as the jargon dejour.
Copyright © 1999 Kevin J. Walsh | [back to Powers of Two] |
walsh@njit.edu /KJW | disclaimer |