ex7.yml

# the alpha parameter
alpha: 0.125
# global data-size (excluding the number of ghost layers for boundary conditions)
global_size: { height: 60, width: 12 }
# degree of parallelism (number of blocks in each dimension)
parallelism: { height: 2, width: 2 }
# PDI configuration
pdi:
  metadata: # small values for which PDI keeps a copy
    ii: int
    dsize: { type: array, subtype: int, size: 2 }
    psize: { type: array, subtype: int, size: 2 }
    pcoord: { type: array, subtype: int, size: 2 }
  data: # values for which PDI does not keep a copy
    main_field: { type: array, subtype: double, size: [ '$dsize[0]', '$dsize[1]' ] }
  plugins:
    decl_hdf5:
      file: ex7-data-${pcoord[0]}x${pcoord[1]}_iter_${ii}.h5
      datasets: # a list of datasets inside the file created on first access
        #*** define the datasets for main_field
        #...
      write:
        main_field:
          when: '$ii=1' # we can use a field specific condition
          #dataset: 'main_field' # no need to specify the dataset name if it is equal to the PDI data name
          #(ex10 presents an example with a different variable name)
          #*** use memory_selection to eliminate the ghost layer
          memory_selection:
            size: [..., ...] # number of elements to transfer in each dimension
            start: [..., ...] # coordinate of the start point in memory relative to the shared data