Re: 32-bit BIGINT value

[Rick James]

A 32-bit system has to work a little hard to simulate 64-bit arithmetic. Before computers expanded from 16-bit to 32-bit, there was an issue of doing 32-bit arithmetic. (That was in the '90s.) Before that was the jump from 8-bit to 16-bit. There were even a few microprocessors in the '70s that were "4-bit".

Will we every go to 128? Already, under the covers, there are datapaths that are bigger than 64-bit. But there is not much need for 128-bit arithmetic in the foreseeable future. Sure, there are always a few situations where big integers are needed, and there is no limit on what someone will need.

64 is enough to do just about anything by count the atoms in the universe.

IP addresses are moving from 32-bit to 128-bit, but you don't "add" them.

MD5 is 128 bits; SHA-1 is 160 bits. Those do need arithmetic to calculate, but it is very isolated.

IEEE-754 Floating point standard (which is used by virtually every computer today) has some 106 (or so) bit temporary results.



Edited 1 time(s). Last edit at 01/11/2011 09:26PM by Rick James.