Re: [Boost-users] [mpl] compile time integer power

From: Sami (randomaccessiterator_at_[hidden])
Date: 2007-01-22 18:51:27

On 1/22/07, Sami <randomaccessiterator_at_[hidden]> wrote:
>
> Hi Tony,
>
> Thx for your reply first of all.
>
> On 1/22/07, Gottlob Frege <gottlobfrege_at_[hidden]> wrote:
> >
> >
> >
> > On 1/22/07, Sami <randomaccessiterator_at_[hidden] > wrote:
> > >
> > > Hi All,
> > >
> > >
> > >
> >     if (Pow == 0) return 1; // to do argue about the value of '0 to the
> > 0'.
> >     if (Pow == 1) return r;
> >     return pow<Pow - 1>(r) * r;
> >
>
>  }
> >
> >
> Now, why do you want it - I doubt that it will buy you anything in terms
> of performance.
> The issue is not particularly about performance,  it is about being able
> to do this at compile time; other than, well, trying out MPL. Your
> resolution to the problem is quite nice, though that's not what I am asking.
> My question is about MPL, not "how to do this with C++"...
>
> There should be a way of doing this using MPL with a single line of code.
> Or, even compute integer power of integers and get a compile time result.
> This is useful for many metaprograms and again, not a performance concern.
>
> Any ideas?
>
>
>

One more note. In my original post, I mentioned this constant sequence
wrapper. If there were such a thing (is there?) then one could evaluate
integer power of an integer at compile time. For instance:

const unsigned SIZE = 3;
const unsigned EIGHT  = fold ( constant_sequence< int_<2>, SIZE >,
int_<1>, times< _ , _ > :: type :: value ;

This is only a thought. Maybe there is a better way of computing powers,
without using fold or that constant_sequence thingy..

As for the real powers, maybe there is a Boost.Fusion algorithm that can do
it..

sami.