int- > float precision?

[Kram1032]

http://ompf.org/forum/viewtopic.php?f=4&t=1209
sounds pretty interesting to me....

[alex22]

Another example: If you take an int64 variable and count up from 0 by 1 every microsecond, it will take 213,5 Million years before it reaches max.If you do the same thing with int32, it takes just a bit over an hour.
If you should use int256 (which is not supported by hardware), you would reach precision on time scale that is already unreal. Counting up by 1 every time the shortest time-span our physical theories can describe today, the Planck-time, the variable would be full after a time that is 10^22 times the age of the universe. Now thats enough time to describe about anything.

[Kram1032]

Nice example

Well, a high-precision integer is probably easier to hack together and make it efficient enough, than a highprecision float... or?