[WIP] Multigrid support

[landinjm]

Do not review. Posted to collaborate with @fractalsbyx.

Summary of issues that are closed:
#394, #398. #386, #376, #409, #393, #294, #329, #313, #276, #290, #311

Issues in progress
#351
#346
#323
#301
#291
#285
#273
#253
#218
#210
#174
#90
#86
#65

[landinjm]

We can just integrate the work in #379 as we rebuild computeRHS, computeLHS & postprocess.

[landinjm]

VariableValueContainer can be deleted too

[landinjm]

Checklist of features left to integrate

• AMR
• Newton update & nonlinear solves
• Multiple equations
• User-facing implementations
• Initial conditions
• Boundary conditions
• Grain growth features
• User-controlled inputs
• File read-in
• Checkpoints
• Transient problems
• Nucleation

[landinjm]

For 1D support it would nice to not change any potential indexing for vector fields when switching to 1D. In the get_vector_* commands we should add something to avoid switching return types:

  using vectorValue = typename std::conditional<
    dim == 1,
    dealii::VectorizedArray<number>,
    dealii::Tensor<1, dim, dealii::VectorizedArray<number>>>::type;
  using vectorGrad = typename std::conditional<
    dim == 1,
    dealii::Tensor<1, dim, dealii::VectorizedArray<number>>,
    dealii::Tensor<2, dim, dealii::VectorizedArray<number>>>::type;
  using vectorHess = typename std::conditional<
    dim == 1,
    dealii::Tensor<2, dim, dealii::VectorizedArray<number>>,
    dealii::Tensor<3, dim, dealii::VectorizedArray<number>>>::type;

This is done

[landinjm]

For the new variableAttributes we should test the most efficient method of lookup (see here) since the evaluation flags are passed for every cell.

[landinjm]

Should add more units tests for variableAttributes now that there's more working parts