TensorLayout

class dolfin.cpp.la.TensorLayout(*args)

Bases: dolfin.cpp.common.Variable

This class described the size and possibly the sparsity of a (sparse) tensor. It is used by the linear algebra backends to initialise tensors.

Create a tensor layout.

Parameters:
  • mpi_comm (MPI_Comm) –
  • std::shared_ptr< const IndexMap >> index_maps (std::vector<) –
  • primary_dim (std::size_t) –
  • sparsity_pattern (Sparsity) –
  • ghosted (Ghosts) –
Ghosts_GHOSTED = True
Ghosts_UNGHOSTED = False
Sparsity_DENSE = False
Sparsity_SPARSE = True
index_map()

Return IndexMap for dimension.

Parameters:i (std::size_t) –
Return type:std::shared_ptr< const IndexMap >
init()

Initialize tensor layout.

Parameters:
  • std::shared_ptr< const IndexMap >> index_maps (std::vector<) –
  • ghosted (Ghosts) –
Return type:

void
is_ghosted()

Require ghosts.

Return type:Ghosts
local_range()

Return local range for dimension dim.

Parameters:dim (std::size_t) –
Return type:std::pair< std::size_t, std::size_t >
mpi_comm()

Return MPI communicator.

Return type:MPI_Comm
primary_dim

Primary storage dim (e.g., 0=row major, 1=column major)

Return type:const std::size_t
rank()

Return rank.

Return type:std::size_t
size()

Return global size for dimension i (size of tensor, includes non-zeroes)

Parameters:i (std::size_t) –
Return type:std::size_t
sparsity_pattern()

Return sparsity pattern (possibly null)

Return type:std::shared_ptr< SparsityPattern >
thisown

The membership flag