Todo List

Module lagrange_interpolation

• Improve efficiency of sll_m_lagrange_interpolation_1d.F90
• Implement other interpolation routines.

Module maxwell_solvers

Implement sll_maxwell_3d_fdtd

Add examples in documentation

Module nml_mesh_cart

add initialization in 2D,3D,4D if needed

Module particle_groups

implement control variates

Module particle_mesh_coupling

Implement 3d splines and other basis functions, e.g. Fourier

Subprogram sll_m_collective::sll_collective_bcast_comp32 (col, buffer, size, root)

WHY DO WE NEED IT?

Subprogram sll_m_collective::sll_collective_bcast_comp64 (col, buffer, size, root)

WHY DO WE NEED IT?

Subprogram sll_m_collective::sll_collective_bcast_real32 (col, buffer, size, root)

WHY DO WE NEED IT?

Subprogram sll_m_collective::sll_s_collective_bcast_3d_real64 (col, buffer, root)

WHY DO WE NEED IT?

Subprogram sll_m_collective::sll_s_collective_bcast_real64 (col, buffer, size, root)

WHY DO WE NEED IT?

Subprogram sll_m_gnuplot_parallel::sll_s_gnuplot_rect_2d_parallel (x_min, delta_x, y_min, delta_y, npts_x, npts_y, array, array_name, iplot, error)

There is a problem with this function. This function should be explicitly told which collective to use, thus permitting the writing of data by a subset of the processors if needed. This assumes that all the processors will contain the writable data. The collective should be passed as an argument. But this may imply that different collectives will write data with the same name... further changes are needed.

Module sll_m_hdf5_serial

Add detailed description

Module sll_m_hermite_interpolator_2d

for the moment only in implementation for the case DIRICHLET x PERIODIC

Module sll_m_interpolators_1d_base

delete function for this type

Module sll_m_interpolators_2d_base

Function names should be reviewed and improved. What is the best way to express that a derivative is in a particular direction? Why eta???

Module sll_m_io_utilities

Add detailed description

Subprogram sll_m_linear_operator_block::set_linear_operator_block (self, i_block_row, i_block_col, linop, r_coeff)

use linop % n_rows/cols instead of n_global_rows/cols

Subprogram sll_m_linear_operator_kron::sll_transpose (x, x_t)

to be optimized depending on the biggest size

Subprogram sll_m_linear_solver_block::set_linear_solver_block (self, i_block_row, i_block_col, linop, r_coeff)

use linop % n_rows/cols instead of n_global_rows/cols

Subprogram sll_m_linear_solver_iter_abstract::set_linop_linear_solver_iter_abstract (self, linear_operator)

add nullspace arguments

Subprogram sll_m_linear_solver_kron::solve_3_real_linear_solver_kron (self, rhs, unknown)

Module sll_m_matrix_csr

must use is_allocated in the create of csr matrices (probably we won't create the graph etc ...)

Module sll_m_ode_integrator_base

Add detailed description

Module sll_m_pic_poisson_base

Try to integrate also general elliptic and quasi neutral solvers.

Allow for more than one weights

Functions to compute derivatives

Module sll_m_sparse_grid_2d

Implement the optimized interpolation routines for option boundary=1

Module sll_m_sparse_grid_3d

Implement the optimized interpolation routines for option boundary=1

Module sll_m_xdmf_light_parallel

Add detailed description

Module sll_m_xdmf_light_serial

Add detailed description

Module sll_m_xml

Add detailed description

Module sparse_grid

Improve unit tests.