Re: min/max by value - 25% faster

I see a ~25% increase in performance (VC2010 SP1) when adding an
overload that takes integers by value instead of const references.
Why are these not implemented as a part of STL for int, double, etc..?

#define MIN(a, b) a<b?a:b
#define MAX(a, b) a>b?a:b

namespace std
{
        inline const int min(int const a, int const b)
        {
                return MIN(a, b);
        }
        inline const int max(int const a, int const b)
        {
                return MAX(a, b);
        }

}

caller:
{
        std::vector<int> v(2);
        v[0] = get_some_random_number();
        v[1] = get_some_random_number();
        int m = std::min(v[0], v[1]); // ~25% faster when the o=

is chosen

}