Lees-Edwards sweep. More...

Public Member Functions
	InterpolateAndShiftAtBoundary (std::shared_ptr< StructuredBlockForest > blocks, BlockDataID field_id, BlockDataID tmp_field_id, unsigned int n_ghost_layers, unsigned int shear_direction, unsigned int shear_plane_normal, std::function< double()> get_pos_offset, std::function< double()> get_shift=[]() { return 0.0;})

FloatType	get_pos_offset () const

FloatType	get_shift () const

void	operator() (IBlock *block)

Detailed Description

template<class FieldType, typename FloatType>
class walberla::InterpolateAndShiftAtBoundary< FieldType, FloatType >

Lees-Edwards sweep.

Todo:: Currently only works for 1 MPI rank! It should work in parallel if the MPI domain decomposition for the structured block forest doesn't partition along the shear direction. For example if the shear direction goes along the z-axis, it should be possible to run on 4 MPI ranks with [2, 2, 1]. At the moment, ESPResSo requires system.cell_system.node_grid to be in decreasing order, therefore parallelization requires a shear direction along the z-axis and a MPI node_grid of [x, y, 1] with x >= y. This restriction on the ordering of the node_grid may be lifted in the distant future, when our FFT algorithm is replaced by a new one.

Constructor & Destructor Documentation

template<class FieldType , typename FloatType >

walberla::InterpolateAndShiftAtBoundary< FieldType, FloatType >::InterpolateAndShiftAtBoundary	(	std::shared_ptr< StructuredBlockForest >	blocks,
		BlockDataID	field_id,
		BlockDataID	tmp_field_id,
		unsigned int	n_ghost_layers,
		unsigned int	shear_direction,
		unsigned int	shear_plane_normal,
		std::function< double()>	get_pos_offset,
		std::function< double()>	get_shift = `[]() { return 0.0; }`
	)