offload/test/offloading/fortran/dump_map_tables.f90

   1 ! Offloading test with runtine call to ompx_dump_mapping_tables
   2 ! Fortran array writing some values and printing the variable mapped to device
   3 ! correctly receives the updates made on the device.
   4 ! REQUIRES: flang
   5 ! UNSUPPORTED: nvptx64-nvidia-cuda-LTO
   6 ! UNSUPPORTED: aarch64-unknown-linux-gnu
   7 ! UNSUPPORTED: aarch64-unknown-linux-gnu-LTO
   8 ! UNSUPPORTED: x86_64-pc-linux-gnu
   9 ! UNSUPPORTED: x86_64-pc-linux-gnu-LTO
  10
  11 ! RUN: %libomptarget-compile-fortran-run-and-check-generic
  12
  13 program map_dump_example
  14   INTERFACE
  15     SUBROUTINE ompx_dump_mapping_tables() BIND(C)
  16     END SUBROUTINE ompx_dump_mapping_tables
  17   END INTERFACE
  18
  19   integer i,j,k,N
  20   integer async_q(4)
  21   real :: A(5000000)
  22   N=5000000
  23   do i=1, N
  24     A(i)=0
  25   enddo
  26 ! clang-format off
  27 ! CHECK: omptarget device 0 info: OpenMP Host-Device pointer mappings after block
  28 ! CHECK-NEXT: omptarget device 0 info: Host Ptr Target Ptr Size (B) DynRefCount HoldRefCount Declaration
  29 ! CHECK-NEXT: omptarget device 0 info: {{(0x[0-9a-f]{16})}} {{(0x[0-9a-f]{16})}}  20000000 1 0 {{.*}} at a(:n):21:11
  30 ! clang-format on
  31 !$omp target enter data map(to:A(:N))
  32   call ompx_dump_mapping_tables()
  33 !$omp target parallel do
  34   do i=1, N
  35     A(i)=A(i)*2
  36   enddo
  37 !$omp target exit data map(from:A)
  38 end program