the legal status of x64 productions
category: general [glöplog]
i wonder... are all the people saying "why bother using 64bit at all?" still using 16bit on the same basis? Anyway, my thoughts on it: if you're using 64bit stuff, you're getting some advantage from it presumably. If bigger code is the downside, it balances out. No need for 4.6k compos.
the switch from 16->32bit gave you a whole bunch of new addressing modes (including the ability to use all general purpose registers for adressing), lots of new instructions, and a flat memory model.
64bit gives you 16 general purpose registers, which is neat, and a prefix byte for nearly every instruction, which isn't.
as for 64bit integer arithmetic... how much of that do you have in current programs? now compare that to the amount of 32bit integer arithmetic you had in 16bit programs.
32bit is just enough in most practical applications.
64bit gives you 16 general purpose registers, which is neat, and a prefix byte for nearly every instruction, which isn't.
as for 64bit integer arithmetic... how much of that do you have in current programs? now compare that to the amount of 32bit integer arithmetic you had in 16bit programs.
32bit is just enough in most practical applications.
it's nice for fixed point arithmetic, 32bit int is a pain there :)
just for products. and when you're doing fixedpoint with 64bit integers, you want 128bit intermediates in products ;)
i got some routines which could be _much_ faster with 64bit adress space (sparse addressing rules), yawn i should buy a quad-g5 mac before they actually retire...
