> Oops, boo-boo.
>
> I was a little over-general in the int/float cpu comment.
> Not all arithmetic operations are equal. What I was getting
> at is the division specifically. With coprocessor on some
> machines (my Alpha at least) floating divs are faster.
>
> Melvin
>
Well this is a problem for code being moved from one platform to annother.
If the programmer (on an alpha say) knows that they take a performance hit
from interger maths, because the floting point stuff has been optimized to
the exclusion of the integer maths ops (it's a workstation, maths is one
of it's things ;), then they don't use it, then when the code is moved
over to i386 (say), where floating point maths is slower than the
corresponding interger op, all the unnecessary (or avoidable) fp maths has
to be rewritten.
Bryn
--
PGP key pass phrase forgotten, \ Overload -- core meltdown sequence
again :( | initiated.
/ This space is intentionally left
| blank, apart from this text ;-)
\____________________________________