Adding 4d+5d versions of copy dp to sp

src/mo_fortran_tools.F90

@@ -167,6 +167,8 @@ MODULE mo_fortran_tools
     MODULE PROCEDURE copy_5d_sp
     MODULE PROCEDURE copy_2d_dpsp
     MODULE PROCEDURE copy_3d_dpsp
+    MODULE PROCEDURE copy_4d_dpsp
+    MODULE PROCEDURE copy_5d_dpsp
     MODULE PROCEDURE copy_2d_spdp
     MODULE PROCEDURE copy_3d_spdp
     MODULE PROCEDURE copy_4d_spdp
@@ -912,6 +914,81 @@ CONTAINS
 
   END SUBROUTINE copy_3d_dpsp
 
+  !> copy state, omp parallel, does not wait for other threads to complete
+  SUBROUTINE copy_4d_dpsp(src, dest, lacc, opt_acc_async)
+    REAL(dp), INTENT(IN) :: src(:, :, :, :)
+    REAL(sp), INTENT(OUT) :: dest(:, :, :, :)
+    LOGICAL, INTENT(IN) :: lacc
+    LOGICAL, INTENT(IN), OPTIONAL :: opt_acc_async
+    INTEGER :: i1, i2, i3, i4, m1, m2, m3, m4
+    LOGICAL :: lzacc
+
+    CALL set_acc_host_or_device(lzacc, lacc)
+
+    m1 = SIZE(dest, 1)
+    m2 = SIZE(dest, 2)
+    m3 = SIZE(dest, 3)
+    m4 = SIZE(dest, 4)
+
+    !$ACC PARALLEL LOOP DEFAULT(PRESENT) ASYNC(1) COLLAPSE(4) IF(lzacc)
+#if (defined(__INTEL_COMPILER))
+!$omp do private(i1,i2,i3,i4)
+#else
+!$omp do collapse(4)
+#endif
+    DO i4 = 1, m4
+      DO i3 = 1, m3
+        DO i2 = 1, m2
+          DO i1 = 1, m1
+            dest(i1, i2, i3, i4) = REAL(src(i1, i2, i3, i4), KIND=sp)
+          END DO
+        END DO
+      END DO
+    END DO
+!$omp end do nowait
+    CALL acc_wait_if_requested(1, opt_acc_async)
+
+  END SUBROUTINE copy_4d_dpsp
+
+  !> copy state, omp parallel, does not wait for other threads to complete
+  SUBROUTINE copy_5d_dpsp(src, dest, lacc, opt_acc_async)
+    REAL(dp), INTENT(IN) :: src(:, :, :, :, :)
+    REAL(sp), INTENT(OUT) :: dest(:, :, :, :, :)
+    LOGICAL, INTENT(IN) :: lacc
+    LOGICAL, INTENT(IN), OPTIONAL :: opt_acc_async
+    INTEGER :: i1, i2, i3, i4, i5, m1, m2, m3, m4, m5
+    LOGICAL :: lzacc
+
+    CALL set_acc_host_or_device(lzacc, lacc)
+
+    m1 = SIZE(dest, 1)
+    m2 = SIZE(dest, 2)
+    m3 = SIZE(dest, 3)
+    m4 = SIZE(dest, 4)
+    m5 = SIZE(dest, 5)
+
+    !$ACC PARALLEL LOOP DEFAULT(PRESENT) ASYNC(1) COLLAPSE(5) IF(lzacc)
+#if (defined(__INTEL_COMPILER))
+!$omp do private(i1,i2,i3,i4,i5)
+#else
+!$omp do collapse(5)
+#endif
+    DO i5 = 1, m5
+      DO i4 = 1, m4
+        DO i3 = 1, m3
+          DO i2 = 1, m2
+            DO i1 = 1, m1
+              dest(i1, i2, i3, i4, i5) = REAL(src(i1, i2, i3, i4, i5), KIND=sp)
+            END DO
+          END DO
+        END DO
+      END DO
+    END DO
+!$omp end do nowait
+
+    CALL acc_wait_if_requested(1, opt_acc_async)
+  END SUBROUTINE copy_5d_dpsp
+
   !> copy state, omp parallel, does not wait for other threads to complete
   SUBROUTINE copy_2d_spdp(src, dest, lacc, opt_acc_async)
     REAL(sp), INTENT(IN) :: src(:, :)