.. _ShellMatrix_Constructors:

############
Constructors
############


The :ref:`ShellMatrix<ShellMatrix>` should be returned from the :ref:`Operator class<Operator>` via :ref:`create_ShellMatrix(...)<Operator_Functions_Frontends_create_ShellMatrix>`.
All member functions are ``const`` can **not** be modified after creation.
We choose this design to ensure the correct communication patter when :ref:`MPI<MPI_docs>` is used.
Hence, the constructors are ``protected`` and can by only used by the :ref:`Operator class<Operator>`.
If :ref:`MPI<MPI_docs>` is **not** enable, it is constructed by


.. doxygenfunction:: danceq::internal::ShellMatrix::ShellMatrix(const Op* operator_ptr_)
 
Otherwise, the following constructor (with :ref:`MPI<MPI_docs>`) is used:

.. note::

    :ref:`MPI<MPI_docs>` requires additional memory for the communication between its processes.
    This can by limited by **overhead_in_GB_per_core** in :ref:`Operator class<Operator>` via :ref:`create_ShellMatrix(...)<Operator_Functions_Frontends_create_ShellMatrix>`.


.. doxygenfunction:: danceq::internal::ShellMatrix::ShellMatrix(const Op* operator_ptr_, const int32_t myrank_, const int32_t world_size_, const uint64_t start_, const uint64_t end_, const uint64_t number_of_communication_steps_, const uint64_t maximal_number_of_elements_to_recv_, const uint64_t maximal_number_of_elements_to_send_, const std::vector<uint64_t> dim_per_step_, const std::vector<uint64_t> total_number_of_elements_to_recv_per_step_, const std::vector<uint64_t> total_number_of_elements_to_send_per_step_, const std::vector<uint64_t> ownership_per_rank_, const std::vector<std::vector<uint64_t>> number_of_elements_to_send_per_rank_per_step_, const std::vector<std::vector<uint64_t>> number_of_elements_to_recv_per_rank_per_step_)