Calculate the evaluation and jacobians of geometry mappings of cells
using sum factorization. By avoiding all geometry mapping of intersections
we can avoid the face-twist problem and just make the grid consistens (not
yet done).

TODO:
- Make grid consistent
- Vectorization
- 3D

Not yet implemented:

- Vectorization of quadrature loop
- Anisotropic amount of quadrature points per direction

René Heß authored 6 years ago and Dominic Kempf committed 6 years ago

Rename the function and remove the backend switch. Reasoning:

- This function was only called once without the backend switching
  mechanism.
- The function takes different arguments than the pdelab
  pymbolic_quadrature_position.
- It returns only one component of the quadrature point.

René Heß authored 6 years ago and Dominic Kempf committed 6 years ago

Probably not the solution we want to have in the long run but right
now it is not possible to have a backend switch for
default/blockstructured/sumfact. This way the old code should behave
the same way as before and it is possible to switch between those
three things.

Note: This is still structured for now

Rename the function and remove the backend switch. Reasoning:

- This function was only called once without the backend switching
  mechanism.
- The function takes different arguments than the pdelab
  pymbolic_quadrature_position.
- It returns only one component of the quadrature point.

Probably not the solution we want to have in the long run but right
now it is not possible to have a backend switch for
default/blockstructured/sumfact. This way the old code should behave
the same way as before and it is possible to switch between those
three things.

This will leave the coverage of the testing suite untouched!

They are only to a lesser extent important for our HPC applications
and the vectorization algorithm struggles with the complexity of the
search space.

So far, we had the opcounter cast to Dune::FieldVector
of size 1. Now it defines the correct type.