Explicit Replication Using Templates

If an array is not explicitly distributed it is given the default mapping - often, but not always, replication.

To force replication of an array, A:

• declare a template to be the same size as the processor grid,
• align A with each template element,
• distribute the template

For example,

      REAL, DIMENSION(100,100)   :: A
!HPF$ PROCESSORS, DIMENSION(NUMBER_OF_PROCESSORS()) :: P
!HPF$ TEMPLATE, DIMENSION(NUMBER_OF_PROCESSORS())   :: T
!HPF$ ALIGN A(*,*) WITH T(*)
!HPF$ DISTRIBUTE (BLOCK) :: T

For more information, click here