Go to the source code of this file.
|
| type | distributed_oct_m::distributed_t |
| | Distribution of N instances over mpi_grpsize processes, for the local rank mpi_grprank. Distribution for all other processes of mpi_grpcomm are stored in process, num and range. More...
|
| |
|
| pure integer function | distributed_oct_m::distributed_nlocal (this) |
| | Number of instances of the distributed quantity on process mpi_grprank. More...
|
| |
| pure integer function | distributed_oct_m::distributed_nglobal (this) |
| | Total number of distributed instances. More...
|
| |
| pure integer function, dimension(:), allocatable | distributed_oct_m::distributed_displs (this) |
| | Displacement of each process''s local part, relative to the start of the distributed quantity, as required by MPI (all)gatherv and (all)scatterv. More...
|
| |
| subroutine, public | distributed_oct_m::distributed_init (this, total, comm, tag, scalapack_compat) |
| | Distribute N instances across M processes of communicator comm More...
|
| |
| subroutine, public | distributed_oct_m::distributed_init_serial (this, total) |
| | Serial initialization of a distributed instance. The calling process is assigned all total instances, such that start = 1, end = total and parallel = .false., and no MPI communication is required. More...
|
| |
| subroutine, public | distributed_oct_m::distributed_copy (in, out) |
| | Create a copy of a distributed instance. More...
|
| |
| subroutine, public | distributed_oct_m::distributed_end (this) |
| |