Really easy C++ programming question

[Erica]

How does one represent n! in C++?

There's got to be a more direct way than creating a loop.

If it's relavant to the answer, I am creating templates for calculating probability density functions. I am aware that I can calculate these in Excel using built in functions but what fun would that be?

I looked in math.h, but couldn't find what I was looking for.

Thanks in advance.

 

[Adam]

Erica,

To my knowledge, there is no factorial function or operator native to C++. If you want to compute a factorial, you'll have to write a function for it yourself. You can do this pretty easily with a loop or, if you're really determined to make life difficult for yourself, a recursive function.

 

[alain]

Erica,

To my knowledge, there is no factorial function or operator native to C++. If you want to compute a factorial, you'll have to write a function for it yourself. You can do this pretty easily with a loop or, if you're really determined to make life difficult for yourself, a recursive function.

Actually the recursive function is the easiest way to do it.

double fact(const double& n)
{
    if (n <= 1.) return 1.;
    return n * factorial(n - 1.);
}

The main complain about this implementation is that it could overflow but any decent compiler will unwind the tail recursion.

 

[SteveTownsend]

You can use template metaprogramming to speed up the computation of factorials. This is advanced C++, though.

C++ Template Metaprogramming has an example.

If you are programming a more conventional factorial function, don't recurse or your stack may overflow. Use a loop. And make sure you don't overflow the int bounds and wind up with a garbage value. n! gets very large very quickly.

 

[alain]

If you are programming a more conventional factorial function, don't recurse or your stack may overflow.

Indeed, the template version should be the best solution and, as I said before, any decent compiler should unwind the tail recursion and should not overflow the stack.

 

[Erica]

thanks all

I'l post my solution soon.

This community's sophisticated level of knowledge and absolute willingness to help never ceases to amaze me. Thanks again.

Have a great day and enjoy the dancing banana
:dance:

 

[SteveTownsend]

factorial overflow

using unsigned int (32 bits) this demonstrates how few factorials can be calculated before you overflow the supported range of numbers:

0! = 1
5! = 120
10! = 3628800
15! = 2004310016
18! = 3396534272
19! = 109641728


The template solution is good for wow! factor but not very practical in my view since it does not alow you to do:

unsigned int n(5);
unsigned int nfactorial = factorial<n>::value;

You have to know n at compile-time. Nice for constants that you use a lot, not for general usage. perhaps there is a way round this but I don't have time to look into it now.