|
MFC
Exascale flow solver
|
|
MFC
Exascale flow solver
|
MPI communication layer: domain decomposition, halo exchange, reductions, and parallel I/O setup. More...
Functions/Subroutines | |
| impure subroutine | s_initialize_mpi_common_module (exchange_all_chemistry_temperatures_in, use_rdma_transport_in) |
| Initialize the module. | |
| impure subroutine | s_mpi_initialize |
| Initialize the MPI execution environment and query the number of processors and local rank. | |
| impure subroutine | s_initialize_mpi_data (q_cons_vf, ib_markers, ib_mpi_data, beta, qbmm_pb, qbmm_mv) |
| Set up MPI I/O data views and variable pointers for parallel file output. | |
| subroutine | s_initialize_mpi_data_ds (m_ds, n_ds, p_ds, q_cons_vf) |
| Set up MPI I/O data views for downsampled (coarsened) parallel file output. | |
| impure subroutine | s_mpi_gather_data (my_vector, counts, gathered_vector, root) |
| Gather variable-length real vectors from all MPI ranks onto the root process. | |
| impure subroutine | mpi_bcast_time_step_values (proc_time, time_avg) |
| Gather per-rank time step wall-clock times onto rank 0 for performance reporting. | |
| impure subroutine | s_prohibit_abort (condition, message) |
| Print a case file error with the prohibited condition and message, then abort execution. | |
| impure subroutine | s_mpi_reduce_stability_criteria_extrema (icfl_max_loc, vcfl_max_loc, rc_min_loc, bubs_loc, icfl_max_glb, vcfl_max_glb, rc_min_glb, bubs_glb, ccfl_max_loc, ccfl_max_glb) |
| The goal of this subroutine is to determine the global extrema of the stability criteria in the computational domain. This is performed by sifting through the local extrema of each stability criterion. Note that each of the local extrema is from a single process, within its assigned section of the computational domain. Finally, note that the global extrema values are only bookkeept on the rank 0 processor. | |
| subroutine | s_mpi_reduce_int_sum (var_loc, sum) |
| Reduce a local integer value to its global sum across all MPI ranks. | |
| impure subroutine | s_mpi_allreduce_sum (var_loc, var_glb) |
| Reduce a local real value to its global sum across all MPI ranks. | |
| impure subroutine | s_mpi_allreduce_vectors_sum (var_loc, var_glb, num_vectors, vector_length) |
| Reduce an array of vectors to their global sums across all MPI ranks. | |
| impure subroutine | s_mpi_allreduce_integer_sum (var_loc, var_glb) |
| Reduce a local integer value to its global sum across all MPI ranks. | |
| impure subroutine | s_mpi_allreduce_min (var_loc, var_glb) |
| Reduce a local real value to its global minimum across all MPI ranks. | |
| impure subroutine | s_mpi_allreduce_max (var_loc, var_glb) |
| Reduce a local real value to its global maximum across all MPI ranks. | |
| impure subroutine | s_mpi_reduce_min (var_loc) |
| Reduce a local real value to its global minimum across all ranks. | |
| impure subroutine | s_mpi_reduce_maxloc (var_loc) |
| Reduce a 2-element variable to its global maximum value with the owning processor rank (MPI_MAXLOC). Reduce a local value to its global maximum with location (rank) across all ranks. | |
| impure subroutine | s_mpi_abort (prnt, code) |
| The subroutine terminates the MPI execution environment. | |
| impure subroutine | s_mpi_barrier |
| Halts all processes until all have reached barrier. | |
| impure subroutine | s_mpi_finalize |
| The subroutine finalizes the MPI execution environment. | |
| subroutine | s_mpi_sendrecv_variables_buffers (q_comm, mpi_dir, pbc_loc, nvar, pb_in, mv_in, q_t_sf) |
| The goal of this procedure is to populate the buffers of the cell-average conservative variables by communicating with the neighboring processors. | |
| subroutine | s_mpi_reduce_beta_variables_buffers (q_comm, kahan_comp, mpi_dir, pbc_loc, nvar) |
| The goal of this procedure is to populate the buffers of the cell-average conservative variables by communicating with the neighboring processors. | |
| subroutine | s_mpi_decompose_computational_domain (write_silo_ghost_offsets, adjust_local_domains, output_offsets, local_domains) |
| The purpose of this procedure is to optimally decompose the computational domain among the available processors. This is performed by attempting to award each processor, in each of the coordinate directions, approximately the same number of cells, and then recomputing the affected global parameters. | |
| subroutine, private | s_apply_decomposition_policies (proc_counts, remainders, local_cells, global_cells, write_silo_ghost_offsets, adjust_local_domains, output_offsets, local_domains) |
| Apply executable-configured output and local-domain policies after the shared Cartesian decomposition. | |
| subroutine | s_mpi_sendrecv_grid_variable_buffer (cell_boundaries, cell_centers, cell_widths, num_cells, bc_bounds, pbc_loc, offset) |
| The goal of this procedure is to populate the buffers of the grid variables by communicating with the neighboring processors. Note that only the buffers of the cell-width distributions are handled in such a way. This is because the buffers of cell-boundary locations may be calculated directly from those of the cell-width distributions. | |
| subroutine | s_apply_grid_from_global_dim (x_cb_glb, m_dim_glb, m_dim, sidx, bc_beg, bc_end, cb_lo, cb_hi, cw_lo, cw_hi, x_cb_loc, x_cc_loc, dx_loc) |
| Populate the local cell-boundary, cell-center, and cell-width arrays in one direction directly from the global cell-boundary array. This guarantees that every rank sees bitwise-identical values at any shared physical cell or boundary. | |
| impure subroutine | s_finalize_mpi_common_module |
| Module deallocation and/or disassociation procedures. | |
Variables | |
| integer, private | v_size |
| real(wp), dimension(:), allocatable, private | buff_send |
| Primitive variable send buffer for halo exchange. | |
| real(wp), dimension(:), allocatable, private | buff_recv |
| Primitive variable receive buffer for halo exchange Variables for EL bubbles communication. | |
| type(int_bounds_info), dimension(3) | comm_coords |
| integer, dimension(3) | comm_size |
| integer, dimension(1:3) | beta_vars = [1, 2, 5] |
| q_beta indices to communicate: 1=void fraction, 2=d(beta)/dt, 5=energy source | |
| integer(kind=8) | halo_size |
| logical, private | exchange_all_chemistry_temperatures = .false. |
| logical, private | use_rdma_transport = .false. |
MPI communication layer: domain decomposition, halo exchange, reductions, and parallel I/O setup.
| impure subroutine m_mpi_common::mpi_bcast_time_step_values | ( | real(wp), dimension(0:num_procs - 1), intent(inout) | proc_time, |
| real(wp), intent(inout) | time_avg ) |
Gather per-rank time step wall-clock times onto rank 0 for performance reporting.
Definition at line 722 of file m_mpi_common.fpp.f90.
|
private |
Apply executable-configured output and local-domain policies after the shared Cartesian decomposition.
Definition at line 3280 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_apply_grid_from_global_dim | ( | real(wp), dimension(-1:m_dim_glb), intent(in) | x_cb_glb, |
| integer, intent(in) | m_dim_glb, | ||
| integer, intent(in) | m_dim, | ||
| integer, intent(in) | sidx, | ||
| integer, intent(in) | bc_beg, | ||
| integer, intent(in) | bc_end, | ||
| integer, intent(in) | cb_lo, | ||
| integer, intent(in) | cb_hi, | ||
| integer, intent(in) | cw_lo, | ||
| integer, intent(in) | cw_hi, | ||
| real(wp), dimension(-1 - cb_lo:m_dim + cb_hi), intent(inout) | x_cb_loc, | ||
| real(wp), dimension(-cw_lo:m_dim + cw_hi), intent(inout) | x_cc_loc, | ||
| real(wp), dimension(-cw_lo:m_dim + cw_hi), intent(inout) | dx_loc ) |
Populate the local cell-boundary, cell-center, and cell-width arrays in one direction directly from the global cell-boundary array. This guarantees that every rank sees bitwise-identical values at any shared physical cell or boundary.
Definition at line 3399 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_finalize_mpi_common_module |
Module deallocation and/or disassociation procedures.
Definition at line 3477 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_initialize_mpi_common_module | ( | logical, intent(in) | exchange_all_chemistry_temperatures_in, |
| logical, intent(in) | use_rdma_transport_in ) |
Initialize the module.
Definition at line 416 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_initialize_mpi_data | ( | type(scalar_field), dimension(sys_size), intent(in) | q_cons_vf, |
| type(integer_field), intent(in), optional | ib_markers, | ||
| type(mpi_io_ib_var), intent(inout), optional | ib_mpi_data, | ||
| type(scalar_field), intent(in), optional | beta, | ||
| type(pres_field), intent(in), optional | qbmm_pb, | ||
| type(pres_field), intent(in), optional | qbmm_mv ) |
Set up MPI I/O data views and variable pointers for parallel file output.
Definition at line 572 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_initialize_mpi_data_ds | ( | integer, intent(in) | m_ds, |
| integer, intent(in) | n_ds, | ||
| integer, intent(in) | p_ds, | ||
| type(scalar_field), dimension(sys_size), intent(in), optional | q_cons_vf ) |
Set up MPI I/O data views for downsampled (coarsened) parallel file output.
Definition at line 655 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_abort | ( | character(len=*), intent(in), optional | prnt, |
| integer, intent(in), optional | code ) |
The subroutine terminates the MPI execution environment.
Definition at line 944 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_allreduce_integer_sum | ( | integer(kind=8), intent(in) | var_loc, |
| integer(kind=8), intent(out) | var_glb ) |
Reduce a local integer value to its global sum across all MPI ranks.
Definition at line 864 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_allreduce_max | ( | real(wp), intent(in) | var_loc, |
| real(wp), intent(out) | var_glb ) |
Reduce a local real value to its global maximum across all MPI ranks.
Definition at line 894 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_allreduce_min | ( | real(wp), intent(in) | var_loc, |
| real(wp), intent(out) | var_glb ) |
Reduce a local real value to its global minimum across all MPI ranks.
Definition at line 880 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_allreduce_sum | ( | real(wp), intent(in) | var_loc, |
| real(wp), intent(out) | var_glb ) |
Reduce a local real value to its global sum across all MPI ranks.
Definition at line 829 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_allreduce_vectors_sum | ( | real(wp), dimension(:,:), intent(in) | var_loc, |
| real(wp), dimension(:,:), intent(inout) | var_glb, | ||
| integer, intent(in) | num_vectors, | ||
| integer, intent(in) | vector_length ) |
Reduce an array of vectors to their global sums across all MPI ranks.
Definition at line 843 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_barrier |
Halts all processes until all have reached barrier.
Definition at line 975 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_mpi_decompose_computational_domain | ( | logical, intent(in) | write_silo_ghost_offsets, |
| logical, intent(in) | adjust_local_domains, | ||
| type(int_bounds_info), dimension(3), intent(inout), optional | output_offsets, | ||
| type(bounds_info), dimension(3), intent(inout), optional | local_domains ) |
The purpose of this procedure is to optimally decompose the computational domain among the available processors. This is performed by attempting to award each processor, in each of the coordinate directions, approximately the same number of cells, and then recomputing the affected global parameters.
Non-optimal number of processors in the x-, y- and z-directions
Definition at line 2886 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_finalize |
The subroutine finalizes the MPI execution environment.
Definition at line 986 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_gather_data | ( | real(wp), dimension(counts), intent(in) | my_vector, |
| integer, intent(in) | counts, | ||
| real(wp), dimension(:), intent(out), allocatable | gathered_vector, | ||
| integer, intent(in) | root ) |
Gather variable-length real vectors from all MPI ranks onto the root process.
| [in] | counts | Array of vector lengths for each process |
| [in] | my_vector | Input vector on each process |
| [in] | root | Rank of the root process |
| [out] | gathered_vector | Gathered vector on the root process |
Definition at line 692 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_initialize |
Initialize the MPI execution environment and query the number of processors and local rank.
Definition at line 537 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_mpi_reduce_beta_variables_buffers | ( | type(scalar_field), dimension(1:), intent(inout) | q_comm, |
| type(scalar_field), dimension(1:), intent(inout) | kahan_comp, | ||
| integer, intent(in) | mpi_dir, | ||
| integer, intent(in) | pbc_loc, | ||
| integer, intent(in) | nvar ) |
The goal of this procedure is to populate the buffers of the cell-average conservative variables by communicating with the neighboring processors.
| q_cons_vf | Cell-average conservative variables |
| mpi_dir | MPI communication coordinate direction |
| pbc_loc | Processor boundary condition (PBC) location |
Definition at line 2346 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_mpi_reduce_int_sum | ( | integer, intent(in) | var_loc, |
| integer, intent(out) | sum ) |
Reduce a local integer value to its global sum across all MPI ranks.
Definition at line 813 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_reduce_maxloc | ( | real(wp), dimension(2), intent(inout) | var_loc | ) |
Reduce a 2-element variable to its global maximum value with the owning processor rank (MPI_MAXLOC). Reduce a local value to its global maximum with location (rank) across all ranks.
Definition at line 927 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_reduce_min | ( | real(wp), intent(inout) | var_loc | ) |
Reduce a local real value to its global minimum across all ranks.
Definition at line 908 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_mpi_reduce_stability_criteria_extrema | ( | real(wp), intent(in) | icfl_max_loc, |
| real(wp), intent(in) | vcfl_max_loc, | ||
| real(wp), intent(in) | rc_min_loc, | ||
| integer, intent(in) | bubs_loc, | ||
| real(wp), intent(out) | icfl_max_glb, | ||
| real(wp), intent(out) | vcfl_max_glb, | ||
| real(wp), intent(out) | rc_min_glb, | ||
| integer, intent(out) | bubs_glb, | ||
| real(wp), intent(in) | ccfl_max_loc, | ||
| real(wp), intent(out) | ccfl_max_glb ) |
The goal of this subroutine is to determine the global extrema of the stability criteria in the computational domain. This is performed by sifting through the local extrema of each stability criterion. Note that each of the local extrema is from a single process, within its assigned section of the computational domain. Finally, note that the global extrema values are only bookkeept on the rank 0 processor.
Definition at line 755 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_mpi_sendrecv_grid_variable_buffer | ( | real(wp), dimension(-1 - offset%beg:), intent(inout), contiguous | cell_boundaries, |
| real(wp), dimension(-buff_size:), intent(inout), contiguous | cell_centers, | ||
| real(wp), dimension(-buff_size:), intent(inout), contiguous | cell_widths, | ||
| integer, intent(in) | num_cells, | ||
| type(int_bounds_info), intent(in) | bc_bounds, | ||
| integer, intent(in) | pbc_loc, | ||
| type(int_bounds_info), intent(in) | offset ) |
The goal of this procedure is to populate the buffers of the grid variables by communicating with the neighboring processors. Note that only the buffers of the cell-width distributions are handled in such a way. This is because the buffers of cell-boundary locations may be calculated directly from those of the cell-width distributions.
Definition at line 3344 of file m_mpi_common.fpp.f90.
| subroutine m_mpi_common::s_mpi_sendrecv_variables_buffers | ( | type(scalar_field), dimension(1:), intent(inout) | q_comm, |
| integer, intent(in) | mpi_dir, | ||
| integer, intent(in) | pbc_loc, | ||
| integer, intent(in) | nvar, | ||
| real(stp), dimension(idwbuff(1)%beg:,idwbuff(2)%beg:,idwbuff(3)%beg:,1:,1:), intent(inout), optional | pb_in, | ||
| real(stp), dimension(idwbuff(1)%beg:,idwbuff(2)%beg:,idwbuff(3)%beg:,1:,1:), intent(inout), optional | mv_in, | ||
| type(scalar_field), intent(inout), optional | q_t_sf ) |
The goal of this procedure is to populate the buffers of the cell-average conservative variables by communicating with the neighboring processors.
Definition at line 998 of file m_mpi_common.fpp.f90.
| impure subroutine m_mpi_common::s_prohibit_abort | ( | character(len=*), intent(in) | condition, |
| character(len=*), intent(in) | message ) |
Print a case file error with the prohibited condition and message, then abort execution.
Definition at line 736 of file m_mpi_common.fpp.f90.
| integer, dimension(1:3) m_mpi_common::beta_vars = [1, 2, 5] |
q_beta indices to communicate: 1=void fraction, 2=d(beta)/dt, 5=energy source
Definition at line 370 of file m_mpi_common.fpp.f90.
|
private |
Primitive variable receive buffer for halo exchange Variables for EL bubbles communication.
Definition at line 366 of file m_mpi_common.fpp.f90.
|
private |
Primitive variable send buffer for halo exchange.
Definition at line 364 of file m_mpi_common.fpp.f90.
| type(int_bounds_info), dimension(3) m_mpi_common::comm_coords |
Definition at line 367 of file m_mpi_common.fpp.f90.
| integer, dimension(3) m_mpi_common::comm_size |
Definition at line 368 of file m_mpi_common.fpp.f90.
|
private |
Definition at line 410 of file m_mpi_common.fpp.f90.
| integer(kind=8) m_mpi_common::halo_size |
Definition at line 397 of file m_mpi_common.fpp.f90.
|
private |
Definition at line 411 of file m_mpi_common.fpp.f90.
|
private |
Definition at line 351 of file m_mpi_common.fpp.f90.