matrix and vector template classes in c++

Question

#include <array>

template <typename T>
class Vector4<T> {
    std::array<T, 4> _a; // or 'T _a[4];'?
};

template <typename T>
class Matrix4<T> {
    std::array<T, 16> _a; // or 'T _a[16];'?
    //Vector4<T> row0; // or should i use this instead
    //Vector4<T> row1; // it makes other code easier but how
    //Vector4<T> row2; // can i implement something like 
    //Vector4<T> row3; // std::array::data()?
};

thanks

edit: ya its for 3d game programming... so i will need more then boost matrix can provide anyway like rotate and translate and invert etc...

Answer 1

I don't think you would want std::array _a[4], as you will create 4 arrays with that.

Answer 2

There's already a standard std::vector class. If you need something specifically for linear algebra, I'd suggest looking into boost ublas.

Answer 3

The way you want to do it, would be

std::vector<std::vector<int> > my_matrix(4, std::vector<int>(4));

However, I would rather use ublas from boost if you want to handle matricies:

#include <boost/numeric/ublas/matrix.hpp>
#include <boost/numeric/ublas/io.hpp>

int main () {
    using namespace boost::numeric::ublas;
    matrix<double> m (3, 3);
    for (unsigned i = 0; i < m.size1 (); ++ i)
        for (unsigned j = 0; j < m.size2 (); ++ j)
            m (i, j) = 3 * i + j;
    std::cout << m << std::endl;
}

The example is from their website http://www.boost.org/doc/libs/1%5F40%5F0/libs/numeric/ublas/doc/matrix.htm

Answer 4

Boost ublas is a great library, but is not optimized if all you want to is 4-D stuff. I would recommend keeping the two classes independent:

template <typename T>
class Vector4<T> {
    T _a[4];
};

template <typename T>
class Matrix4<T> {
    T _m[4][4];
};

You can always cast a pointer to a row (_m[i]) into a Vector4<> if needed. To get a column vector though you will need an adapter class like ublas uses.

Answer 5

I've found Eigen to be the most straightforward of the C++ linear algebra libraries, and it contains templates for fixed and variable dimension vectors and matrices. Like Boost, it's a pure template "library" so there are no libs to build / include, but I find it to be more complete and significantly more performant than Boost's ublas.