Implement transpose

[wolfv]

We should implement a transpose operation to transpose NxN matrix blocks (where N is batch width). The interface should probably look like the one for haddp, e.g. taking a pointer of rows.

template <class T, N>
void transpose(batch<T, N>* rows)
{
... inplace transpose ...
}

[wolfv]

Note MIPP has very readable implementations for this. https://github.com/aff3ct/MIPP

[wolfv]

Also implementations available in PacketMath of Eigen3.