Order of quadrature rules on polytopes

[fdrmrc]

In several examples we use quadrature rules of order $2p+1$ inside each element of the sub-tessellation, which are definitely an overkill, especially in 3D.
I've verified locally that decreasing the order (say to $p+1$ only) does not affect accuracy in our examples and gives much better assembly times thanks to better usage of caches.