pow() function question

Jarrett Billingsley kb3ctd2 at yahoo.com
Tue Nov 27 08:56:52 PST 2007


"aeschere" <aeschere at nospam.cryptobit.org> wrote in message 
news:fiheps$2m28$1 at digitalmars.com...
>I am relatively new to D.  So far, I really like it.  However, I need to
> do some power functions on integers.  Is pow() the only way of
> performing this type of function? It seems silly to have to either use
> reals for integer work.  This was one of my gripes with C's pow()
> function.  I check the news archives and noticed there was talk of
> creating a '**' power function for integers back in 2005. I tried it but
> it didn't work.
>
> Michael
> (remove nospam if you need to get a hold of me personally)

If you're using phobos, std.math.pow has an overload for integral exponents. 
Although I think it's only highly optimized for lower powers, like up to the 
third or fourth; anything higher and I think it uses floating point numbers.

Keep in mind that floating point performance on modern processors is not 
exactly slow.  Doing a power operation (which involves.. what, a ln? 
something like that) might be faster than a loop that multiplies the 
integers over and over. 




More information about the Digitalmars-d-learn mailing list