The test cases have lots of common calls which are shared by many tests. We could generate a test_common.f90 or test_base.f90
to store these common routines for all of the tests. For example this is shared by the tests:
subroutine initialise_mpi()
call MPI_Init(ierr)
call MPI_Comm_rank(MPI_COMM_WORLD, nrank, ierr)
call MPI_Comm_size(MPI_COMM_WORLD, nproc, ierr)
if (nrank == 0) print *, 'Performance benchmark with', nproc, 'ranks'
pnext = modulo(nrank - nproc + 1, nproc)
pprev = modulo(nrank - 1, nproc)
end subroutine initialise_mpi This way we could reduce the complexity of standalone tests
The test cases have lots of common calls which are shared by many tests. We could generate a
test_common.f90ortest_base.f90to store these common routines for all of the tests. For example this is shared by the tests:
This way we could reduce the complexity of standalone tests