Re: Why is using single-precision slower than using double-precision

>The above code verges on inaccuracy by keeping an index variable as
>a floating point number and incrementing it by a floating point
>number. Systematic error is introduced into such an index variable if the


In general that is the case, especially when he constants 0.1 or 0.01 are
used to increment. However in IEEE floating point format, pure integers
add to give pure integers if they don't get too big. So the code shown
would work correctly on a machine with IEEE representation.


Roedy@bix.com
-30-
--

Return to the comp.compilers page.
Search the comp.compilers archives again.