Fine to coarse grid filters

Implement fine to coarse grid filters with remeshing kernels:

• Python implementation

Implement fine to coarse grid filters with FFT:

• Python implementation
• OpenCL implementation
• Implement energy computation on all backends
• Implement energy plotter and energy dumper

Compatibility, testing and examples:

• Deprecate MultiresolutionFilter and move it to the common interface, fix the corresponding test.
• Add an example using the new operator: examples/multiresolution/scalar_advection.py
• Add examples using energy dumpers and plotters:
  1. examples/multiresolution/scalar_advection.py
  2. examples/taylor_green/taylor_green.py
  3. examples/particles_above_salt/particles_above_salt_bc_3d.py

New features:

• Added MultiComputationalGraphNodeFrontend to enable method dependent ComputationalGraphNodeFrontends.
• Better graph support for multigrid problems.
• Better debug logs for FieldRequirements and GraphBuilder.
• New symbolic kernel generation features:
  • Vectorization is now disabled by default for the symbolic code generator.
  • Added support for indexed buffer assignment
  • Added support for complex modulus and complex squared modulus.
  • Added support for round and cast.
  • Variables can now be force to be declared volatile which allows the implementation of critical section by using local or global mutexes.
  • Added support for CodeSections.

Bug fixes:

• Directional advection autotuning used a wrong timestep resulting in INVALID_COMMAND_QUEUE on the first run.
• 2D multiscale advection offset fix in generated opencl code.