| ! Test lower of elemental user defined assignments |
| ! RUN: bbc -emit-fir -hlfir=false %s -o - | FileCheck %s |
| |
| module defined_assignments |
| type t |
| integer :: i |
| end type |
| interface assignment(=) |
| elemental subroutine assign_t(a,b) |
| import t |
| type(t),intent(out) :: a |
| type(t),intent(in) :: b |
| end |
| end interface |
| interface assignment(=) |
| elemental subroutine assign_logical_to_real(a,b) |
| real, intent(out) :: a |
| logical, intent(in) :: b |
| end |
| end interface |
| interface assignment(=) |
| elemental subroutine assign_real_to_logical(a,b) |
| logical, intent(out) :: a |
| real, intent(in) :: b |
| end |
| end interface |
| end module |
| |
| ! CHECK-LABEL: func @_QPtest_derived( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>> {fir.bindc_name = "x"}) { |
| ! CHECK: %[[C_100:[-0-9a-z_]+]] = arith.constant 100 : index |
| ! CHECK: %[[V_0:[0-9]+]] = fir.shape %[[C_100]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_1:[0-9]+]] = fir.array_load %arg0(%[[V_0]]) : (!fir.ref<!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>>, !fir.shape<1>) -> !fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>> |
| ! CHECK: %[[C_100_i64:[-0-9a-z_]+]] = arith.constant 100 : i64 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_100_i64]] : (i64) -> index |
| ! CHECK: %[[C_m1_i64:[-0-9a-z_]+]] = arith.constant -1 : i64 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_m1_i64]] : (i64) -> index |
| ! CHECK: %[[C_1_i64:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_4:[0-9]+]] = fir.convert %[[C_1_i64]] : (i64) -> index |
| ! CHECK: %[[V_5:[0-9]+]] = fir.shape %[[C_100]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.slice %[[V_2]], %[[V_4]], %[[V_3:[0-9]+]] : (index, index, index) -> !fir.slice<1> |
| ! CHECK: %[[V_7:[0-9]+]] = fir.array_load %arg0(%[[V_5]]) [%[[V_6]]] : (!fir.ref<!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>>, !fir.shape<1>, !fir.slice<1>) -> !fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>> |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_8:[0-9]+]] = arith.subi %[[C_100]], %[[C_1]] : index |
| ! CHECK: %[[V_9:[0-9]+]] = fir.do_loop %arg1 = %[[C_0]] to %[[V_8]] step %[[C_1]] unordered iter_args(%arg2 = %[[V_1]]) -> (!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>) { |
| ! CHECK: %[[V_10:[0-9]+]] = fir.array_access %[[V_7]], %arg1 : (!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>, index) -> !fir.ref<!fir.type<_QMdefined_assignmentsTt{i:i32}>> |
| ! CHECK: %[[V_11:[0-9]+]] = fir.no_reassoc %[[V_10:[0-9]+]] : !fir.ref<!fir.type<_QMdefined_assignmentsTt{i:i32}>> |
| ! CHECK: %[[V_12:[0-9]+]]:2 = fir.array_modify %arg2, %arg1 : (!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>, index) -> (!fir.ref<!fir.type<_QMdefined_assignmentsTt{i:i32}>>, !fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>) |
| ! CHECK: fir.call @_QPassign_t(%[[V_12]]#0, %[[V_11]]) fastmath<contract> : (!fir.ref<!fir.type<_QMdefined_assignmentsTt{i:i32}>>, !fir.ref<!fir.type<_QMdefined_assignmentsTt{i:i32}>>) -> () |
| ! CHECK: fir.result %[[V_12]]#1 : !fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_1]], %[[V_9]] to %arg0 : !fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>, !fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>, !fir.ref<!fir.array<100x!fir.type<_QMdefined_assignmentsTt{i:i32}>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_intrinsic( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "x"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca !fir.logical<4> |
| ! CHECK: %[[C_100:[-0-9a-z_]+]] = arith.constant 100 : index |
| ! CHECK: %[[V_1:[0-9]+]] = fir.shape %[[C_100]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_2:[0-9]+]] = fir.array_load %arg0(%[[V_1]]) : (!fir.ref<!fir.array<100xf32>>, !fir.shape<1>) -> !fir.array<100xf32> |
| ! CHECK: %[[C_100_i64:[-0-9a-z_]+]] = arith.constant 100 : i64 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_100_i64]] : (i64) -> index |
| ! CHECK: %[[C_m1_i64:[-0-9a-z_]+]] = arith.constant -1 : i64 |
| ! CHECK: %[[V_4:[0-9]+]] = fir.convert %[[C_m1_i64]] : (i64) -> index |
| ! CHECK: %[[C_1_i64:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_5:[0-9]+]] = fir.convert %[[C_1_i64]] : (i64) -> index |
| ! CHECK: %[[V_6:[0-9]+]] = fir.shape %[[C_100]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_7:[0-9]+]] = fir.slice %[[V_3]], %[[V_5]], %[[V_4:[0-9]+]] : (index, index, index) -> !fir.slice<1> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.array_load %arg0(%[[V_6]]) [%[[V_7]]] : (!fir.ref<!fir.array<100xf32>>, !fir.shape<1>, !fir.slice<1>) -> !fir.array<100xf32> |
| ! CHECK: %[[C_st:[-0-9a-z_]+]] = arith.constant 0.000000e+00 : f32 |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_9:[0-9]+]] = arith.subi %[[C_100]], %[[C_1]] : index |
| ! CHECK: %[[V_10:[0-9]+]] = fir.do_loop %arg1 = %[[C_0]] to %[[V_9]] step %[[C_1]] unordered iter_args(%arg2 = %[[V_2]]) -> (!fir.array<100xf32>) { |
| ! CHECK: %[[V_11:[0-9]+]] = fir.array_fetch %[[V_8]], %arg1 : (!fir.array<100xf32>, index) -> f32 |
| ! CHECK: %[[V_12:[0-9]+]] = arith.cmpf olt, %[[V_11]], %[[C_st]] {{.*}} : f32 |
| ! CHECK: %[[V_13:[0-9]+]]:2 = fir.array_modify %arg2, %arg1 : (!fir.array<100xf32>, index) -> (!fir.ref<f32>, !fir.array<100xf32>) |
| ! CHECK: %[[V_14:[0-9]+]] = fir.convert %[[V_12:[0-9]+]] : (i1) -> !fir.logical<4> |
| ! CHECK: fir.store %[[V_14]] to %[[V_0:[0-9]+]] : !fir.ref<!fir.logical<4>> |
| ! CHECK: fir.call @_QPassign_logical_to_real(%[[V_13]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<f32>, !fir.ref<!fir.logical<4>>) -> () |
| ! CHECK: fir.result %[[V_13]]#1 : !fir.array<100xf32> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_2]], %[[V_10]] to %arg0 : !fir.array<100xf32>, !fir.array<100xf32>, !fir.ref<!fir.array<100xf32>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_intrinsic_2( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<100x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "y"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca f32 |
| ! CHECK: %[[C_100:[-0-9a-z_]+]] = arith.constant 100 : index |
| ! CHECK: %[[C_100_0:[-0-9a-z_]+]] = arith.constant 100 : index |
| ! CHECK: %[[V_1:[0-9]+]] = fir.shape %[[C_100]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_2:[0-9]+]] = fir.array_load %arg0(%[[V_1]]) : (!fir.ref<!fir.array<100x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<100x!fir.logical<4>> |
| ! CHECK: %[[V_3:[0-9]+]] = fir.shape %[[C_100_0]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_4:[0-9]+]] = fir.array_load %arg1(%[[V_3]]) : (!fir.ref<!fir.array<100xf32>>, !fir.shape<1>) -> !fir.array<100xf32> |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_5:[0-9]+]] = arith.subi %[[C_100]], %[[C_1]] : index |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[C_0]] to %[[V_5]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_2]]) -> (!fir.array<100x!fir.logical<4>>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.array_fetch %[[V_4]], %arg2 : (!fir.array<100xf32>, index) -> f32 |
| ! CHECK: %[[V_8:[0-9]+]] = fir.no_reassoc %[[V_7:[0-9]+]] : f32 |
| ! CHECK: %[[V_9:[0-9]+]]:2 = fir.array_modify %arg3, %arg2 : (!fir.array<100x!fir.logical<4>>, index) -> (!fir.ref<!fir.logical<4>>, !fir.array<100x!fir.logical<4>>) |
| ! CHECK: fir.store %[[V_8]] to %[[V_0:[0-9]+]] : !fir.ref<f32> |
| ! CHECK: fir.call @_QPassign_real_to_logical(%[[V_9]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<!fir.logical<4>>, !fir.ref<f32>) -> () |
| ! CHECK: fir.result %[[V_9]]#1 : !fir.array<100x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_2]], %[[V_6]] to %arg0 : !fir.array<100x!fir.logical<4>>, !fir.array<100x!fir.logical<4>>, !fir.ref<!fir.array<100x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPfrom_char( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "i"}, %arg1: !fir.box<!fir.array<?x!fir.char<1,?>>> {fir.bindc_name = "c"}) { |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_0:[0-9]+]]:3 = fir.box_dims %arg0, %[[C_0]] : (!fir.box<!fir.array<?xi32>>, index) -> (index, index, index) |
| ! CHECK: %[[V_1:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?xi32>>) -> !fir.array<?xi32> |
| ! CHECK: %[[V_2:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> !fir.array<?x!fir.char<1,?>> |
| ! CHECK: %[[V_3:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = arith.divsi %[[V_3]], %[[C_1]] : index |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_1:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_5:[0-9]+]] = arith.subi %[[V_0]]#1, %[[C_1_0]] : index |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[C_0_1]] to %[[V_5]] step %[[C_1_0]] unordered iter_args(%arg3 = %[[V_1]]) -> (!fir.array<?xi32>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.array_access %[[V_2]], %arg2 typeparams %[[V_4:[0-9]+]] : (!fir.array<?x!fir.char<1,?>>, index, index) -> !fir.ref<!fir.char<1,?>> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[V_9:[0-9]+]] = fir.no_reassoc %[[V_7:[0-9]+]] : !fir.ref<!fir.char<1,?>> |
| ! CHECK: %[[V_10:[0-9]+]]:2 = fir.array_modify %arg3, %arg2 : (!fir.array<?xi32>, index) -> (!fir.ref<i32>, !fir.array<?xi32>) |
| ! CHECK: %[[V_11:[0-9]+]] = fir.emboxchar %[[V_9]], %[[V_8:[0-9]+]] : (!fir.ref<!fir.char<1,?>>, index) -> !fir.boxchar<1> |
| ! CHECK: fir.call @_QPsfrom_char(%[[V_10]]#0, %[[V_11]]) fastmath<contract> : (!fir.ref<i32>, !fir.boxchar<1>) -> () |
| ! CHECK: fir.result %[[V_10]]#1 : !fir.array<?xi32> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_1]], %[[V_6]] to %arg0 : !fir.array<?xi32>, !fir.array<?xi32>, !fir.box<!fir.array<?xi32>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPto_char( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "i"}, %arg1: !fir.box<!fir.array<?x!fir.char<1,?>>> {fir.bindc_name = "c"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca i32 |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_1:[0-9]+]]:3 = fir.box_dims %arg1, %[[C_0]] : (!fir.box<!fir.array<?x!fir.char<1,?>>>, index) -> (index, index, index) |
| ! CHECK: %[[V_2:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> !fir.array<?x!fir.char<1,?>> |
| ! CHECK: %[[V_3:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?xi32>>) -> !fir.array<?xi32> |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_4:[0-9]+]] = arith.subi %[[V_1]]#1, %[[C_1]] : index |
| ! CHECK: %[[V_5:[0-9]+]] = fir.do_loop %arg2 = %[[C_0_0]] to %[[V_4]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_2]]) -> (!fir.array<?x!fir.char<1,?>>) { |
| ! CHECK: %[[V_6:[0-9]+]] = fir.array_fetch %[[V_3]], %arg2 : (!fir.array<?xi32>, index) -> i32 |
| ! CHECK: %[[V_7:[0-9]+]] = fir.no_reassoc %[[V_6:[0-9]+]] : i32 |
| ! CHECK: %[[V_8:[0-9]+]]:2 = fir.array_modify %arg3, %arg2 : (!fir.array<?x!fir.char<1,?>>, index) -> (!fir.ref<!fir.char<1,?>>, !fir.array<?x!fir.char<1,?>>) |
| ! CHECK: %[[V_9:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[V_10:[0-9]+]] = fir.emboxchar %[[V_8]]#0, %[[V_9:[0-9]+]] : (!fir.ref<!fir.char<1,?>>, index) -> !fir.boxchar<1> |
| ! CHECK: fir.store %[[V_7]] to %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: fir.call @_QPsto_char(%[[V_10]], %[[V_0]]) fastmath<contract> : (!fir.boxchar<1>, !fir.ref<i32>) -> () |
| ! CHECK: fir.result %[[V_8]]#1 : !fir.array<?x!fir.char<1,?>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_2]], %[[V_5]] to %arg1 : !fir.array<?x!fir.char<1,?>>, !fir.array<?x!fir.char<1,?>>, !fir.box<!fir.array<?x!fir.char<1,?>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| subroutine test_derived(x) |
| use defined_assignments |
| type(t) :: x(100) |
| x = x(100:1:-1) |
| end subroutine |
| |
| subroutine test_intrinsic(x) |
| use defined_assignments |
| real :: x(100) |
| x = x(100:1:-1) .lt. 0. |
| end subroutine |
| |
| subroutine test_intrinsic_2(x, y) |
| use defined_assignments |
| logical :: x(100) |
| real :: y(100) |
| x = y |
| end subroutine |
| |
| subroutine from_char(i, c) |
| interface assignment(=) |
| elemental subroutine sfrom_char(a,b) |
| integer, intent(out) :: a |
| character(*),intent(in) :: b |
| end subroutine |
| end interface |
| integer :: i(:) |
| character(*) :: c(:) |
| i = c |
| end subroutine |
| |
| subroutine to_char(i, c) |
| interface assignment(=) |
| elemental subroutine sto_char(a,b) |
| character(*), intent(out) :: a |
| integer,intent(in) :: b |
| end subroutine |
| end interface |
| integer :: i(:) |
| character(*) :: c(:) |
| c = i |
| end subroutine |
| |
| ! ----------------------------------------------------------------------------- |
| ! Test user defined assignments inside FORALL and WHERE |
| ! ----------------------------------------------------------------------------- |
| |
| ! CHECK-LABEL: func @_QPtest_in_forall_1( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<10xf32>> {fir.bindc_name = "y"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca f32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "i"} |
| ! CHECK: %[[C_10:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_10_0:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.shape %[[C_10]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg0(%[[V_4]]) : (!fir.ref<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.shape %[[C_10_0]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_7:[0-9]+]] = fir.array_load %arg1(%[[V_6]]) : (!fir.ref<!fir.array<10xf32>>, !fir.shape<1>) -> !fir.array<10xf32> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_5]]) -> (!fir.array<10x!fir.logical<4>>) { |
| ! CHECK: %[[V_9:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_9]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_10:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_11:[0-9]+]] = fir.convert %[[V_10:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_12:[0-9]+]] = fir.convert %[[V_11:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_13:[0-9]+]] = arith.subi %[[V_12]], %[[C_1_1]] : index |
| ! CHECK: %[[V_14:[0-9]+]] = fir.array_fetch %[[V_7]], %[[V_13:[0-9]+]] : (!fir.array<10xf32>, index) -> f32 |
| ! CHECK: %[[V_15:[0-9]+]] = fir.no_reassoc %[[V_14:[0-9]+]] : f32 |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_16:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_17:[0-9]+]] = fir.convert %[[V_16:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_18:[0-9]+]] = fir.convert %[[V_17:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_19:[0-9]+]] = arith.subi %[[V_18]], %[[C_1_2]] : index |
| ! CHECK: %[[V_20:[0-9]+]]:2 = fir.array_modify %arg3, %[[V_19:[0-9]+]] : (!fir.array<10x!fir.logical<4>>, index) -> (!fir.ref<!fir.logical<4>>, !fir.array<10x!fir.logical<4>>) |
| ! CHECK: fir.store %[[V_15]] to %[[V_0:[0-9]+]] : !fir.ref<f32> |
| ! CHECK: fir.call @_QPassign_real_to_logical(%[[V_20]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<!fir.logical<4>>, !fir.ref<f32>) -> () |
| ! CHECK: fir.result %[[V_20]]#1 : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_5]], %[[V_8]] to %arg0 : !fir.array<10x!fir.logical<4>>, !fir.array<10x!fir.logical<4>>, !fir.ref<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_in_forall_2( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<10xf32>> {fir.bindc_name = "y"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca !fir.logical<4> |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "i"} |
| ! CHECK: %[[C_10:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.shape %[[C_10]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg1(%[[V_4]]) : (!fir.ref<!fir.array<10xf32>>, !fir.shape<1>) -> !fir.array<10xf32> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_5]]) -> (!fir.array<10xf32>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_7]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_8:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_9:[0-9]+]] = fir.convert %[[V_8:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_10:[0-9]+]] = fir.convert %[[V_9:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_11:[0-9]+]] = arith.subi %[[V_10]], %[[C_1_0]] : index |
| ! CHECK: %[[V_12:[0-9]+]] = fir.array_fetch %[[V_5]], %[[V_11:[0-9]+]] : (!fir.array<10xf32>, index) -> f32 |
| ! CHECK: %[[C_st:[-0-9a-z_]+]] = arith.constant 0.000000e+00 : f32 |
| ! CHECK: %[[V_13:[0-9]+]] = arith.cmpf olt, %[[V_12]], %[[C_st]] {{.*}} : f32 |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_14:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_15:[0-9]+]] = fir.convert %[[V_14:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_16:[0-9]+]] = fir.convert %[[V_15:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_17:[0-9]+]] = arith.subi %[[V_16]], %[[C_1_1]] : index |
| ! CHECK: %[[V_18:[0-9]+]]:2 = fir.array_modify %arg3, %[[V_17:[0-9]+]] : (!fir.array<10xf32>, index) -> (!fir.ref<f32>, !fir.array<10xf32>) |
| ! CHECK: %[[V_19:[0-9]+]] = fir.convert %[[V_13:[0-9]+]] : (i1) -> !fir.logical<4> |
| ! CHECK: fir.store %[[V_19]] to %[[V_0:[0-9]+]] : !fir.ref<!fir.logical<4>> |
| ! CHECK: fir.call @_QPassign_logical_to_real(%[[V_18]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<f32>, !fir.ref<!fir.logical<4>>) -> () |
| ! CHECK: fir.result %[[V_18]]#1 : !fir.array<10xf32> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_5]], %[[V_6]] to %arg1 : !fir.array<10xf32>, !fir.array<10xf32>, !fir.ref<!fir.array<10xf32>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_intrinsic_where_1( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<10xf32>> {fir.bindc_name = "y"}, %arg2: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "l"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca f32 |
| ! CHECK: %[[C_10:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_10_0:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_10_1:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_10_2:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[V_1:[0-9]+]] = fir.shape %[[C_10]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_2:[0-9]+]] = fir.array_load %arg2(%[[V_1]]) : (!fir.ref<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_3:[0-9]+]] = fir.allocmem !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_4:[0-9]+]] = fir.shape %[[C_10_2]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %[[V_3]](%[[V_4]]) : (!fir.heap<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_6:[0-9]+]] = arith.subi %[[C_10_2]], %[[C_1]] : index |
| ! CHECK: %[[V_7:[0-9]+]] = fir.do_loop %arg3 = %[[C_0]] to %[[V_6]] step %[[C_1]] unordered iter_args(%arg4 = %[[V_5]]) -> (!fir.array<10x!fir.logical<4>>) { |
| ! CHECK: %[[V_15:[0-9]+]] = fir.array_fetch %[[V_2]], %arg3 : (!fir.array<10x!fir.logical<4>>, index) -> !fir.logical<4> |
| ! CHECK: %[[V_16:[0-9]+]] = fir.array_update %arg4, %[[V_15]], %arg3 : (!fir.array<10x!fir.logical<4>>, !fir.logical<4>, index) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: fir.result %[[V_16:[0-9]+]] : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_5]], %[[V_7]] to %[[V_3:[0-9]+]] : !fir.array<10x!fir.logical<4>>, !fir.array<10x!fir.logical<4>>, !fir.heap<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.shape %[[C_10_2]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_9:[0-9]+]] = fir.shape %[[C_10_0]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_10:[0-9]+]] = fir.array_load %arg0(%[[V_9]]) : (!fir.ref<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_11:[0-9]+]] = fir.shape %[[C_10_1]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_12:[0-9]+]] = fir.array_load %arg1(%[[V_11]]) : (!fir.ref<!fir.array<10xf32>>, !fir.shape<1>) -> !fir.array<10xf32> |
| ! CHECK: %[[C_1_3:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_4:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_13:[0-9]+]] = arith.subi %[[C_10_0]], %[[C_1_3]] : index |
| ! CHECK: %[[V_14:[0-9]+]] = fir.do_loop %arg3 = %[[C_0_4]] to %[[V_13]] step %[[C_1_3]] unordered iter_args(%arg4 = %[[V_10]]) -> (!fir.array<10x!fir.logical<4>>) { |
| ! CHECK: %[[C_1_5:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_15:[0-9]+]] = arith.addi %arg3, %[[C_1_5]] : index |
| ! CHECK: %[[V_16:[0-9]+]] = fir.array_coor %[[V_3]](%[[V_8]]) %[[V_15:[0-9]+]] : (!fir.heap<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>, index) -> !fir.ref<!fir.logical<4>> |
| ! CHECK: %[[V_17:[0-9]+]] = fir.load %[[V_16:[0-9]+]] : !fir.ref<!fir.logical<4>> |
| ! CHECK: %[[V_18:[0-9]+]] = fir.convert %[[V_17:[0-9]+]] : (!fir.logical<4>) -> i1 |
| ! CHECK: %[[V_19:[0-9]+]] = fir.if %[[V_18]] -> (!fir.array<10x!fir.logical<4>>) { |
| ! CHECK: %[[V_20:[0-9]+]] = fir.array_fetch %[[V_12]], %arg3 : (!fir.array<10xf32>, index) -> f32 |
| ! CHECK: %[[V_21:[0-9]+]] = fir.no_reassoc %[[V_20:[0-9]+]] : f32 |
| ! CHECK: %[[V_22:[0-9]+]]:2 = fir.array_modify %arg4, %arg3 : (!fir.array<10x!fir.logical<4>>, index) -> (!fir.ref<!fir.logical<4>>, !fir.array<10x!fir.logical<4>>) |
| ! CHECK: fir.store %[[V_21]] to %[[V_0:[0-9]+]] : !fir.ref<f32> |
| ! CHECK: fir.call @_QPassign_real_to_logical(%[[V_22]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<!fir.logical<4>>, !fir.ref<f32>) -> () |
| ! CHECK: fir.result %[[V_22]]#1 : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } else { |
| ! CHECK: fir.result %arg4 : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.result %[[V_19:[0-9]+]] : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_10]], %[[V_14]] to %arg0 : !fir.array<10x!fir.logical<4>>, !fir.array<10x!fir.logical<4>>, !fir.ref<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: fir.freemem %[[V_3:[0-9]+]] : !fir.heap<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_intrinsic_where_2( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<10xf32>> {fir.bindc_name = "y"}, %arg2: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "l"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca !fir.logical<4> |
| ! CHECK: %[[C_10:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_10_0:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_10_1:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[V_1:[0-9]+]] = fir.shape %[[C_10]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_2:[0-9]+]] = fir.array_load %arg2(%[[V_1]]) : (!fir.ref<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_3:[0-9]+]] = fir.allocmem !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_4:[0-9]+]] = fir.shape %[[C_10_1]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %[[V_3]](%[[V_4]]) : (!fir.heap<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_6:[0-9]+]] = arith.subi %[[C_10_1]], %[[C_1]] : index |
| ! CHECK: %[[V_7:[0-9]+]] = fir.do_loop %arg3 = %[[C_0]] to %[[V_6]] step %[[C_1]] unordered iter_args(%arg4 = %[[V_5]]) -> (!fir.array<10x!fir.logical<4>>) { |
| ! CHECK: %[[V_15:[0-9]+]] = fir.array_fetch %[[V_2]], %arg3 : (!fir.array<10x!fir.logical<4>>, index) -> !fir.logical<4> |
| ! CHECK: %[[V_16:[0-9]+]] = fir.array_update %arg4, %[[V_15]], %arg3 : (!fir.array<10x!fir.logical<4>>, !fir.logical<4>, index) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: fir.result %[[V_16:[0-9]+]] : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_5]], %[[V_7]] to %[[V_3:[0-9]+]] : !fir.array<10x!fir.logical<4>>, !fir.array<10x!fir.logical<4>>, !fir.heap<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.shape %[[C_10_1]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_9:[0-9]+]] = fir.shape %[[C_10_0]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_10:[0-9]+]] = fir.array_load %arg1(%[[V_9]]) : (!fir.ref<!fir.array<10xf32>>, !fir.shape<1>) -> !fir.array<10xf32> |
| ! CHECK: %[[V_11:[0-9]+]] = fir.shape %[[C_10_0]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_12:[0-9]+]] = fir.array_load %arg1(%[[V_11]]) : (!fir.ref<!fir.array<10xf32>>, !fir.shape<1>) -> !fir.array<10xf32> |
| ! CHECK: %[[C_st:[-0-9a-z_]+]] = arith.constant 0.000000e+00 : f32 |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_3:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_13:[0-9]+]] = arith.subi %[[C_10_0]], %[[C_1_2]] : index |
| ! CHECK: %[[V_14:[0-9]+]] = fir.do_loop %arg3 = %[[C_0_3]] to %[[V_13]] step %[[C_1_2]] unordered iter_args(%arg4 = %[[V_10]]) -> (!fir.array<10xf32>) { |
| ! CHECK: %[[C_1_4:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_15:[0-9]+]] = arith.addi %arg3, %[[C_1_4]] : index |
| ! CHECK: %[[V_16:[0-9]+]] = fir.array_coor %[[V_3]](%[[V_8]]) %[[V_15:[0-9]+]] : (!fir.heap<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>, index) -> !fir.ref<!fir.logical<4>> |
| ! CHECK: %[[V_17:[0-9]+]] = fir.load %[[V_16:[0-9]+]] : !fir.ref<!fir.logical<4>> |
| ! CHECK: %[[V_18:[0-9]+]] = fir.convert %[[V_17:[0-9]+]] : (!fir.logical<4>) -> i1 |
| ! CHECK: %[[V_19:[0-9]+]] = fir.if %[[V_18]] -> (!fir.array<10xf32>) { |
| ! CHECK: %[[V_20:[0-9]+]] = fir.array_fetch %[[V_12]], %arg3 : (!fir.array<10xf32>, index) -> f32 |
| ! CHECK: %[[V_21:[0-9]+]] = arith.cmpf olt, %[[V_20]], %[[C_st]] {{.*}} : f32 |
| ! CHECK: %[[V_22:[0-9]+]]:2 = fir.array_modify %arg4, %arg3 : (!fir.array<10xf32>, index) -> (!fir.ref<f32>, !fir.array<10xf32>) |
| ! CHECK: %[[V_23:[0-9]+]] = fir.convert %[[V_21:[0-9]+]] : (i1) -> !fir.logical<4> |
| ! CHECK: fir.store %[[V_23]] to %[[V_0:[0-9]+]] : !fir.ref<!fir.logical<4>> |
| ! CHECK: fir.call @_QPassign_logical_to_real(%[[V_22]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<f32>, !fir.ref<!fir.logical<4>>) -> () |
| ! CHECK: fir.result %[[V_22]]#1 : !fir.array<10xf32> |
| ! CHECK: } else { |
| ! CHECK: fir.result %arg4 : !fir.array<10xf32> |
| ! CHECK: } |
| ! CHECK: fir.result %[[V_19:[0-9]+]] : !fir.array<10xf32> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_10]], %[[V_14]] to %arg1 : !fir.array<10xf32>, !fir.array<10xf32>, !fir.ref<!fir.array<10xf32>> |
| ! CHECK: fir.freemem %[[V_3:[0-9]+]] : !fir.heap<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_scalar_func_but_not_elemental( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<100x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "y"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca i32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "i"} |
| ! CHECK: %[[C_100:[-0-9a-z_]+]] = arith.constant 100 : index |
| ! CHECK: %[[C_100_0:[-0-9a-z_]+]] = arith.constant 100 : index |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.shape %[[C_100]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg0(%[[V_4]]) : (!fir.ref<!fir.array<100x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<100x!fir.logical<4>> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.shape %[[C_100_0]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_7:[0-9]+]] = fir.array_load %arg1(%[[V_6]]) : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> !fir.array<100xi32> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_5]]) -> (!fir.array<100x!fir.logical<4>>) { |
| ! CHECK: %[[V_9:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_9]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_10:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_11:[0-9]+]] = fir.convert %[[V_10:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_12:[0-9]+]] = fir.convert %[[V_11:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_13:[0-9]+]] = arith.subi %[[V_12]], %[[C_1_1]] : index |
| ! CHECK: %[[V_14:[0-9]+]] = fir.array_fetch %[[V_7]], %[[V_13:[0-9]+]] : (!fir.array<100xi32>, index) -> i32 |
| ! CHECK: %[[V_15:[0-9]+]] = fir.no_reassoc %[[V_14:[0-9]+]] : i32 |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_16:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_17:[0-9]+]] = fir.convert %[[V_16:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_18:[0-9]+]] = fir.convert %[[V_17:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_19:[0-9]+]] = arith.subi %[[V_18]], %[[C_1_2]] : index |
| ! CHECK: %[[V_20:[0-9]+]]:2 = fir.array_modify %arg3, %[[V_19:[0-9]+]] : (!fir.array<100x!fir.logical<4>>, index) -> (!fir.ref<!fir.logical<4>>, !fir.array<100x!fir.logical<4>>) |
| ! CHECK: fir.store %[[V_15]] to %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: fir.call @_QPassign_integer_to_logical(%[[V_20]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<!fir.logical<4>>, !fir.ref<i32>) -> () |
| ! CHECK: fir.result %[[V_20]]#1 : !fir.array<100x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_5]], %[[V_8]] to %arg0 : !fir.array<100x!fir.logical<4>>, !fir.array<100x!fir.logical<4>>, !fir.ref<!fir.array<100x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPtest_in_forall_with_cleanup( |
| ! CHECK-SAME: %arg0: !fir.ref<!fir.array<10x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.ref<!fir.array<10xf32>> {fir.bindc_name = "y"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca !fir.box<!fir.heap<f32>> {bindc_name = ".result"} |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "i"} |
| ! CHECK: %[[C_10:[-0-9a-z_]+]] = arith.constant 10 : index |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.shape %[[C_10]] : (index) -> !fir.shape<1> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg0(%[[V_4]]) : (!fir.ref<!fir.array<10x!fir.logical<4>>>, !fir.shape<1>) -> !fir.array<10x!fir.logical<4>> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_5]]) -> (!fir.array<10x!fir.logical<4>>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_7]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.call @_QPreturns_alloc(%[[V_1]]) fastmath<contract> : (!fir.ref<i32>) -> !fir.box<!fir.heap<f32>> |
| ! CHECK: fir.save_result %[[V_8]] to %[[V_0:[0-9]+]] : !fir.box<!fir.heap<f32>>, !fir.ref<!fir.box<!fir.heap<f32>>> |
| ! CHECK: %[[V_9:[0-9]+]] = fir.load %[[V_0:[0-9]+]] : !fir.ref<!fir.box<!fir.heap<f32>>> |
| ! CHECK: %[[V_10:[0-9]+]] = fir.box_addr %[[V_9:[0-9]+]] : (!fir.box<!fir.heap<f32>>) -> !fir.heap<f32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_11:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_12:[0-9]+]] = fir.convert %[[V_11:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_13:[0-9]+]] = fir.convert %[[V_12:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_14:[0-9]+]] = arith.subi %[[V_13]], %[[C_1_0]] : index |
| ! CHECK: %[[V_15:[0-9]+]]:2 = fir.array_modify %arg3, %[[V_14:[0-9]+]] : (!fir.array<10x!fir.logical<4>>, index) -> (!fir.ref<!fir.logical<4>>, !fir.array<10x!fir.logical<4>>) |
| ! CHECK: %[[V_16:[0-9]+]] = fir.convert %[[V_10:[0-9]+]] : (!fir.heap<f32>) -> !fir.ref<f32> |
| ! CHECK: fir.call @_QPassign_real_to_logical(%[[V_15]]#0, %[[V_16]]) fastmath<contract> : (!fir.ref<!fir.logical<4>>, !fir.ref<f32>) -> () |
| ! CHECK: %[[V_17:[0-9]+]] = fir.load %[[V_0:[0-9]+]] : !fir.ref<!fir.box<!fir.heap<f32>>> |
| ! CHECK: %[[V_18:[0-9]+]] = fir.box_addr %[[V_17:[0-9]+]] : (!fir.box<!fir.heap<f32>>) -> !fir.heap<f32> |
| ! CHECK: %[[V_19:[0-9]+]] = fir.convert %[[V_18:[0-9]+]] : (!fir.heap<f32>) -> i64 |
| ! CHECK: %[[C_0_i64:[-0-9a-z_]+]] = arith.constant 0 : i64 |
| ! CHECK: %[[V_20:[0-9]+]] = arith.cmpi ne, %[[V_19]], %[[C_0_i64]] : i64 |
| ! CHECK: fir.if %[[V_20]] { |
| ! CHECK: fir.freemem %[[V_18:[0-9]+]] : !fir.heap<f32> |
| ! CHECK: } |
| ! CHECK: fir.result %[[V_15]]#1 : !fir.array<10x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_5]], %[[V_6]] to %arg0 : !fir.array<10x!fir.logical<4>>, !fir.array<10x!fir.logical<4>>, !fir.ref<!fir.array<10x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| subroutine test_in_forall_1(x, y) |
| use defined_assignments |
| logical :: x(10) |
| real :: y(10) |
| forall (i=1:10) x(i) = y(i) |
| end subroutine |
| |
| subroutine test_in_forall_2(x, y) |
| use defined_assignments |
| logical :: x(10) |
| real :: y(10) |
| forall (i=1:10) y(i) = y(i).lt.0. |
| end subroutine |
| |
| subroutine test_intrinsic_where_1(x, y, l) |
| use defined_assignments |
| logical :: x(10), l(10) |
| real :: y(10) |
| where(l) x = y |
| end subroutine |
| |
| subroutine test_intrinsic_where_2(x, y, l) |
| use defined_assignments |
| logical :: x(10), l(10) |
| real :: y(10) |
| where(l) y = y.lt.0. |
| end subroutine |
| |
| subroutine test_scalar_func_but_not_elemental(x, y) |
| interface assignment(=) |
| ! scalar, but not elemental |
| elemental subroutine assign_integer_to_logical(a,b) |
| logical, intent(out) :: a |
| integer, intent(in) :: b |
| end |
| end interface |
| logical :: x(100) |
| integer :: y(100) |
| ! Scalar assignment in forall should be treated just like elemental |
| ! functions. |
| forall(i=1:10) x(i) = y(i) |
| end subroutine |
| |
| subroutine test_in_forall_with_cleanup(x, y) |
| use defined_assignments |
| interface |
| pure function returns_alloc(i) |
| integer, intent(in) :: i |
| real, allocatable :: returns_alloc |
| end function |
| end interface |
| logical :: x(10) |
| real :: y(10) |
| forall (i=1:10) x(i) = returns_alloc(i) |
| end subroutine |
| |
| ! CHECK-LABEL: func @_QPtest_forall_array( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?x?x!fir.logical<4>>> {fir.bindc_name = "x"}, %arg1: !fir.box<!fir.array<?x?xf32>> {fir.bindc_name = "y"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca f32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "i"} |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?x?x!fir.logical<4>>>) -> !fir.array<?x?x!fir.logical<4>> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x?xf32>>) -> !fir.array<?x?xf32> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_4]]) -> (!fir.array<?x?x!fir.logical<4>>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_7]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_8:[0-9]+]]:3 = fir.box_dims %arg0, %[[C_1_0]] : (!fir.box<!fir.array<?x?x!fir.logical<4>>>, index) -> (index, index, index) |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_9:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_10:[0-9]+]] = fir.convert %[[V_9:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_11:[0-9]+]] = fir.convert %[[V_10:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_12:[0-9]+]] = arith.subi %[[V_11]], %[[C_1_1]] : index |
| ! CHECK: %[[C_1_i64:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_13:[0-9]+]] = fir.convert %[[C_1_i64]] : (i64) -> index |
| ! CHECK: %[[V_14:[0-9]+]] = arith.addi %[[C_1_1]], %[[V_8]]#1 : index |
| ! CHECK: %[[V_15:[0-9]+]] = arith.subi %[[V_14]], %[[C_1_1]] : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_16:[0-9]+]] = arith.subi %[[V_15]], %[[C_1_1]] : index |
| ! CHECK: %[[V_17:[0-9]+]] = arith.addi %[[V_16]], %[[V_13:[0-9]+]] : index |
| ! CHECK: %[[V_18:[0-9]+]] = arith.divsi %[[V_17]], %[[V_13:[0-9]+]] : index |
| ! CHECK: %[[V_19:[0-9]+]] = arith.cmpi sgt, %[[V_18]], %[[C_0]] : index |
| ! CHECK: %[[V_20:[0-9]+]] = arith.select %[[V_19]], %[[V_18]], %[[C_0]] : index |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_21:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_22:[0-9]+]] = fir.convert %[[V_21:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_23:[0-9]+]] = fir.convert %[[V_22:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_24:[0-9]+]] = arith.subi %[[V_23]], %[[C_1_2]] : index |
| ! CHECK: %[[C_1_i64_3:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_25:[0-9]+]] = fir.convert %[[C_1_i64_3]] : (i64) -> index |
| ! CHECK: %[[C_1_4:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_5:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_26:[0-9]+]] = arith.subi %[[V_20]], %[[C_1_4]] : index |
| ! CHECK: %[[V_27:[0-9]+]] = fir.do_loop %arg4 = %[[C_0_5]] to %[[V_26]] step %[[C_1_4]] unordered iter_args(%arg5 = %arg3) -> (!fir.array<?x?x!fir.logical<4>>) { |
| ! CHECK: %[[V_28:[0-9]+]] = arith.subi %[[C_1_2]], %[[C_1_2]] : index |
| ! CHECK: %[[V_29:[0-9]+]] = arith.muli %arg4, %[[V_25:[0-9]+]] : index |
| ! CHECK: %[[V_30:[0-9]+]] = arith.addi %[[V_28]], %[[V_29:[0-9]+]] : index |
| ! CHECK: %[[V_31:[0-9]+]] = fir.array_fetch %[[V_5]], %[[V_24]], %[[V_30:[0-9]+]] : (!fir.array<?x?xf32>, index, index) -> f32 |
| ! CHECK: %[[V_32:[0-9]+]] = fir.no_reassoc %[[V_31:[0-9]+]] : f32 |
| ! CHECK: %[[V_33:[0-9]+]] = arith.subi %[[C_1_1]], %[[C_1_1]] : index |
| ! CHECK: %[[V_34:[0-9]+]] = arith.muli %arg4, %[[V_13:[0-9]+]] : index |
| ! CHECK: %[[V_35:[0-9]+]] = arith.addi %[[V_33]], %[[V_34:[0-9]+]] : index |
| ! CHECK: %[[V_36:[0-9]+]]:2 = fir.array_modify %arg5, %[[V_12]], %[[V_35:[0-9]+]] : (!fir.array<?x?x!fir.logical<4>>, index, index) -> (!fir.ref<!fir.logical<4>>, !fir.array<?x?x!fir.logical<4>>) |
| ! CHECK: fir.store %[[V_32]] to %[[V_0:[0-9]+]] : !fir.ref<f32> |
| ! CHECK: fir.call @_QPassign_real_to_logical(%[[V_36]]#0, %[[V_0]]) fastmath<contract> : (!fir.ref<!fir.logical<4>>, !fir.ref<f32>) -> () |
| ! CHECK: fir.result %[[V_36]]#1 : !fir.array<?x?x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.result %[[V_27:[0-9]+]] : !fir.array<?x?x!fir.logical<4>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_4]], %[[V_6]] to %arg0 : !fir.array<?x?x!fir.logical<4>>, !fir.array<?x?x!fir.logical<4>>, !fir.box<!fir.array<?x?x!fir.logical<4>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| subroutine test_forall_array(x, y) |
| use defined_assignments |
| logical :: x(:, :) |
| real :: y(:, :) |
| forall (i=1:10) x(i, :) = y(i, :) |
| end subroutine |
| |
| ! CHECK-LABEL: func @_QPfrom_char_forall( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "i"}, %arg1: !fir.box<!fir.array<?x!fir.char<1,?>>> {fir.bindc_name = "c"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "j"} |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_3:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?xi32>>) -> !fir.array<?xi32> |
| ! CHECK: %[[V_4:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> !fir.array<?x!fir.char<1,?>> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.do_loop %arg2 = %[[V_1]] to %[[V_2]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_3]]) -> (!fir.array<?xi32>) { |
| ! CHECK: %[[V_6:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_6]] to %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_7:[0-9]+]] = fir.load %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_8:[0-9]+]] = fir.convert %[[V_7:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_9:[0-9]+]] = fir.convert %[[V_8:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_10:[0-9]+]] = arith.subi %[[V_9]], %[[C_1_0]] : index |
| ! CHECK: %[[V_11:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_12:[0-9]+]] = arith.divsi %[[V_11]], %[[C_1_1]] : index |
| ! CHECK: %[[V_13:[0-9]+]] = fir.array_access %[[V_4]], %[[V_10]] typeparams %[[V_12:[0-9]+]] : (!fir.array<?x!fir.char<1,?>>, index, index) -> !fir.ref<!fir.char<1,?>> |
| ! CHECK: %[[V_14:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[V_15:[0-9]+]] = fir.no_reassoc %[[V_13:[0-9]+]] : !fir.ref<!fir.char<1,?>> |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_16:[0-9]+]] = fir.load %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_17:[0-9]+]] = fir.convert %[[V_16:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_18:[0-9]+]] = fir.convert %[[V_17:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_19:[0-9]+]] = arith.subi %[[V_18]], %[[C_1_2]] : index |
| ! CHECK: %[[V_20:[0-9]+]]:2 = fir.array_modify %arg3, %[[V_19:[0-9]+]] : (!fir.array<?xi32>, index) -> (!fir.ref<i32>, !fir.array<?xi32>) |
| ! CHECK: %[[V_21:[0-9]+]] = fir.emboxchar %[[V_15]], %[[V_14:[0-9]+]] : (!fir.ref<!fir.char<1,?>>, index) -> !fir.boxchar<1> |
| ! CHECK: fir.call @_QPsfrom_char(%[[V_20]]#0, %[[V_21]]) fastmath<contract> : (!fir.ref<i32>, !fir.boxchar<1>) -> () |
| ! CHECK: fir.result %[[V_20]]#1 : !fir.array<?xi32> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_3]], %[[V_5]] to %arg0 : !fir.array<?xi32>, !fir.array<?xi32>, !fir.box<!fir.array<?xi32>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPto_char_forall( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "i"}, %arg1: !fir.box<!fir.array<?x!fir.char<1,?>>> {fir.bindc_name = "c"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca i32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "j"} |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> !fir.array<?x!fir.char<1,?>> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?xi32>>) -> !fir.array<?xi32> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_4]]) -> (!fir.array<?x!fir.char<1,?>>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_7]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_8:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_9:[0-9]+]] = fir.convert %[[V_8:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_10:[0-9]+]] = fir.convert %[[V_9:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_11:[0-9]+]] = arith.subi %[[V_10]], %[[C_1_0]] : index |
| ! CHECK: %[[V_12:[0-9]+]] = fir.array_fetch %[[V_5]], %[[V_11:[0-9]+]] : (!fir.array<?xi32>, index) -> i32 |
| ! CHECK: %[[V_13:[0-9]+]] = fir.no_reassoc %[[V_12:[0-9]+]] : i32 |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_14:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_15:[0-9]+]] = fir.convert %[[V_14:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_16:[0-9]+]] = fir.convert %[[V_15:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_17:[0-9]+]] = arith.subi %[[V_16]], %[[C_1_1]] : index |
| ! CHECK: %[[V_18:[0-9]+]]:2 = fir.array_modify %arg3, %[[V_17:[0-9]+]] : (!fir.array<?x!fir.char<1,?>>, index) -> (!fir.ref<!fir.char<1,?>>, !fir.array<?x!fir.char<1,?>>) |
| ! CHECK: %[[V_19:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[V_20:[0-9]+]] = fir.emboxchar %[[V_18]]#0, %[[V_19:[0-9]+]] : (!fir.ref<!fir.char<1,?>>, index) -> !fir.boxchar<1> |
| ! CHECK: fir.store %[[V_13]] to %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: fir.call @_QPsto_char(%[[V_20]], %[[V_0]]) fastmath<contract> : (!fir.boxchar<1>, !fir.ref<i32>) -> () |
| ! CHECK: fir.result %[[V_18]]#1 : !fir.array<?x!fir.char<1,?>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_4]], %[[V_6]] to %arg1 : !fir.array<?x!fir.char<1,?>>, !fir.array<?x!fir.char<1,?>>, !fir.box<!fir.array<?x!fir.char<1,?>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| subroutine from_char_forall(i, c) |
| interface assignment(=) |
| elemental subroutine sfrom_char(a,b) |
| integer, intent(out) :: a |
| character(*),intent(in) :: b |
| end subroutine |
| end interface |
| integer :: i(:) |
| character(*) :: c(:) |
| forall (j=1:10) i(j) = c(j) |
| end subroutine |
| |
| subroutine to_char_forall(i, c) |
| interface assignment(=) |
| elemental subroutine sto_char(a,b) |
| character(*), intent(out) :: a |
| integer,intent(in) :: b |
| end subroutine |
| end interface |
| integer :: i(:) |
| character(*) :: c(:) |
| forall (j=1:10) c(j) = i(j) |
| end subroutine |
| |
| ! CHECK-LABEL: func @_QPfrom_char_forall_array( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?x?xi32>> {fir.bindc_name = "i"}, %arg1: !fir.box<!fir.array<?x?x!fir.char<1,?>>> {fir.bindc_name = "c"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "j"} |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_3:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?x?xi32>>) -> !fir.array<?x?xi32> |
| ! CHECK: %[[V_4:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x?x!fir.char<1,?>>>) -> !fir.array<?x?x!fir.char<1,?>> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.do_loop %arg2 = %[[V_1]] to %[[V_2]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_3]]) -> (!fir.array<?x?xi32>) { |
| ! CHECK: %[[V_6:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_6]] to %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_7:[0-9]+]]:3 = fir.box_dims %arg0, %[[C_1_0]] : (!fir.box<!fir.array<?x?xi32>>, index) -> (index, index, index) |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_8:[0-9]+]] = fir.load %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_9:[0-9]+]] = fir.convert %[[V_8:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_10:[0-9]+]] = fir.convert %[[V_9:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_11:[0-9]+]] = arith.subi %[[V_10]], %[[C_1_1]] : index |
| ! CHECK: %[[C_1_i64:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_12:[0-9]+]] = fir.convert %[[C_1_i64]] : (i64) -> index |
| ! CHECK: %[[V_13:[0-9]+]] = arith.addi %[[C_1_1]], %[[V_7]]#1 : index |
| ! CHECK: %[[V_14:[0-9]+]] = arith.subi %[[V_13]], %[[C_1_1]] : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_15:[0-9]+]] = arith.subi %[[V_14]], %[[C_1_1]] : index |
| ! CHECK: %[[V_16:[0-9]+]] = arith.addi %[[V_15]], %[[V_12:[0-9]+]] : index |
| ! CHECK: %[[V_17:[0-9]+]] = arith.divsi %[[V_16]], %[[V_12:[0-9]+]] : index |
| ! CHECK: %[[V_18:[0-9]+]] = arith.cmpi sgt, %[[V_17]], %[[C_0]] : index |
| ! CHECK: %[[V_19:[0-9]+]] = arith.select %[[V_18]], %[[V_17]], %[[C_0]] : index |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_20:[0-9]+]] = fir.load %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_21:[0-9]+]] = fir.convert %[[V_20:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_22:[0-9]+]] = fir.convert %[[V_21:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_23:[0-9]+]] = arith.subi %[[V_22]], %[[C_1_2]] : index |
| ! CHECK: %[[C_1_i64_3:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_24:[0-9]+]] = fir.convert %[[C_1_i64_3]] : (i64) -> index |
| ! CHECK: %[[C_1_4:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_5:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_25:[0-9]+]] = arith.subi %[[V_19]], %[[C_1_4]] : index |
| ! CHECK: %[[V_26:[0-9]+]] = fir.do_loop %arg4 = %[[C_0_5]] to %[[V_25]] step %[[C_1_4]] unordered iter_args(%arg5 = %arg3) -> (!fir.array<?x?xi32>) { |
| ! CHECK: %[[V_27:[0-9]+]] = arith.subi %[[C_1_2]], %[[C_1_2]] : index |
| ! CHECK: %[[V_28:[0-9]+]] = arith.muli %arg4, %[[V_24:[0-9]+]] : index |
| ! CHECK: %[[V_29:[0-9]+]] = arith.addi %[[V_27]], %[[V_28:[0-9]+]] : index |
| ! CHECK: %[[V_30:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[C_1_6:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_31:[0-9]+]] = arith.divsi %[[V_30]], %[[C_1_6]] : index |
| ! CHECK: %[[V_32:[0-9]+]] = fir.array_access %[[V_4]], %[[V_23]], %[[V_29]] typeparams %[[V_31:[0-9]+]] : (!fir.array<?x?x!fir.char<1,?>>, index, index, index) -> !fir.ref<!fir.char<1,?>> |
| ! CHECK: %[[V_33:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[V_34:[0-9]+]] = fir.no_reassoc %[[V_32:[0-9]+]] : !fir.ref<!fir.char<1,?>> |
| ! CHECK: %[[V_35:[0-9]+]] = arith.subi %[[C_1_1]], %[[C_1_1]] : index |
| ! CHECK: %[[V_36:[0-9]+]] = arith.muli %arg4, %[[V_12:[0-9]+]] : index |
| ! CHECK: %[[V_37:[0-9]+]] = arith.addi %[[V_35]], %[[V_36:[0-9]+]] : index |
| ! CHECK: %[[V_38:[0-9]+]]:2 = fir.array_modify %arg5, %[[V_11]], %[[V_37:[0-9]+]] : (!fir.array<?x?xi32>, index, index) -> (!fir.ref<i32>, !fir.array<?x?xi32>) |
| ! CHECK: %[[V_39:[0-9]+]] = fir.emboxchar %[[V_34]], %[[V_33:[0-9]+]] : (!fir.ref<!fir.char<1,?>>, index) -> !fir.boxchar<1> |
| ! CHECK: fir.call @_QPsfrom_char(%[[V_38]]#0, %[[V_39]]) fastmath<contract> : (!fir.ref<i32>, !fir.boxchar<1>) -> () |
| ! CHECK: fir.result %[[V_38]]#1 : !fir.array<?x?xi32> |
| ! CHECK: } |
| ! CHECK: fir.result %[[V_26:[0-9]+]] : !fir.array<?x?xi32> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_3]], %[[V_5]] to %arg0 : !fir.array<?x?xi32>, !fir.array<?x?xi32>, !fir.box<!fir.array<?x?xi32>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| ! CHECK-LABEL: func @_QPto_char_forall_array( |
| ! CHECK-SAME: %arg0: !fir.box<!fir.array<?x?xi32>> {fir.bindc_name = "i"}, %arg1: !fir.box<!fir.array<?x?x!fir.char<1,?>>> {fir.bindc_name = "c"}) { |
| ! CHECK: %[[V_0:[0-9]+]] = fir.alloca i32 |
| ! CHECK: %[[V_1:[0-9]+]] = fir.alloca i32 {adapt.valuebyref, bindc_name = "j"} |
| ! CHECK: %[[C_1_i32:[-0-9a-z_]+]] = arith.constant 1 : i32 |
| ! CHECK: %[[V_2:[0-9]+]] = fir.convert %[[C_1_i32]] : (i32) -> index |
| ! CHECK: %[[C_10_i32:[-0-9a-z_]+]] = arith.constant 10 : i32 |
| ! CHECK: %[[V_3:[0-9]+]] = fir.convert %[[C_10_i32]] : (i32) -> index |
| ! CHECK: %[[C_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_4:[0-9]+]] = fir.array_load %arg1 : (!fir.box<!fir.array<?x?x!fir.char<1,?>>>) -> !fir.array<?x?x!fir.char<1,?>> |
| ! CHECK: %[[V_5:[0-9]+]] = fir.array_load %arg0 : (!fir.box<!fir.array<?x?xi32>>) -> !fir.array<?x?xi32> |
| ! CHECK: %[[V_6:[0-9]+]] = fir.do_loop %arg2 = %[[V_2]] to %[[V_3]] step %[[C_1]] unordered iter_args(%arg3 = %[[V_4]]) -> (!fir.array<?x?x!fir.char<1,?>>) { |
| ! CHECK: %[[V_7:[0-9]+]] = fir.convert %arg2 : (index) -> i32 |
| ! CHECK: fir.store %[[V_7]] to %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[C_1_0:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_8:[0-9]+]]:3 = fir.box_dims %arg1, %[[C_1_0]] : (!fir.box<!fir.array<?x?x!fir.char<1,?>>>, index) -> (index, index, index) |
| ! CHECK: %[[C_1_1:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_9:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_10:[0-9]+]] = fir.convert %[[V_9:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_11:[0-9]+]] = fir.convert %[[V_10:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_12:[0-9]+]] = arith.subi %[[V_11]], %[[C_1_1]] : index |
| ! CHECK: %[[C_1_i64:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_13:[0-9]+]] = fir.convert %[[C_1_i64]] : (i64) -> index |
| ! CHECK: %[[V_14:[0-9]+]] = arith.addi %[[C_1_1]], %[[V_8]]#1 : index |
| ! CHECK: %[[V_15:[0-9]+]] = arith.subi %[[V_14]], %[[C_1_1]] : index |
| ! CHECK: %[[C_0:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_16:[0-9]+]] = arith.subi %[[V_15]], %[[C_1_1]] : index |
| ! CHECK: %[[V_17:[0-9]+]] = arith.addi %[[V_16]], %[[V_13:[0-9]+]] : index |
| ! CHECK: %[[V_18:[0-9]+]] = arith.divsi %[[V_17]], %[[V_13:[0-9]+]] : index |
| ! CHECK: %[[V_19:[0-9]+]] = arith.cmpi sgt, %[[V_18]], %[[C_0]] : index |
| ! CHECK: %[[V_20:[0-9]+]] = arith.select %[[V_19]], %[[V_18]], %[[C_0]] : index |
| ! CHECK: %[[C_1_2:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[V_21:[0-9]+]] = fir.load %[[V_1:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: %[[V_22:[0-9]+]] = fir.convert %[[V_21:[0-9]+]] : (i32) -> i64 |
| ! CHECK: %[[V_23:[0-9]+]] = fir.convert %[[V_22:[0-9]+]] : (i64) -> index |
| ! CHECK: %[[V_24:[0-9]+]] = arith.subi %[[V_23]], %[[C_1_2]] : index |
| ! CHECK: %[[C_1_i64_3:[-0-9a-z_]+]] = arith.constant 1 : i64 |
| ! CHECK: %[[V_25:[0-9]+]] = fir.convert %[[C_1_i64_3]] : (i64) -> index |
| ! CHECK: %[[C_1_4:[-0-9a-z_]+]] = arith.constant 1 : index |
| ! CHECK: %[[C_0_5:[-0-9a-z_]+]] = arith.constant 0 : index |
| ! CHECK: %[[V_26:[0-9]+]] = arith.subi %[[V_20]], %[[C_1_4]] : index |
| ! CHECK: %[[V_27:[0-9]+]] = fir.do_loop %arg4 = %[[C_0_5]] to %[[V_26]] step %[[C_1_4]] unordered iter_args(%arg5 = %arg3) -> (!fir.array<?x?x!fir.char<1,?>>) { |
| ! CHECK: %[[V_28:[0-9]+]] = arith.subi %[[C_1_2]], %[[C_1_2]] : index |
| ! CHECK: %[[V_29:[0-9]+]] = arith.muli %arg4, %[[V_25:[0-9]+]] : index |
| ! CHECK: %[[V_30:[0-9]+]] = arith.addi %[[V_28]], %[[V_29:[0-9]+]] : index |
| ! CHECK: %[[V_31:[0-9]+]] = fir.array_fetch %[[V_5]], %[[V_24]], %[[V_30:[0-9]+]] : (!fir.array<?x?xi32>, index, index) -> i32 |
| ! CHECK: %[[V_32:[0-9]+]] = fir.no_reassoc %[[V_31:[0-9]+]] : i32 |
| ! CHECK: %[[V_33:[0-9]+]] = arith.subi %[[C_1_1]], %[[C_1_1]] : index |
| ! CHECK: %[[V_34:[0-9]+]] = arith.muli %arg4, %[[V_13:[0-9]+]] : index |
| ! CHECK: %[[V_35:[0-9]+]] = arith.addi %[[V_33]], %[[V_34:[0-9]+]] : index |
| ! CHECK: %[[V_36:[0-9]+]]:2 = fir.array_modify %arg5, %[[V_12]], %[[V_35:[0-9]+]] : (!fir.array<?x?x!fir.char<1,?>>, index, index) -> (!fir.ref<!fir.char<1,?>>, !fir.array<?x?x!fir.char<1,?>>) |
| ! CHECK: %[[V_37:[0-9]+]] = fir.box_elesize %arg1 : (!fir.box<!fir.array<?x?x!fir.char<1,?>>>) -> index |
| ! CHECK: %[[V_38:[0-9]+]] = fir.emboxchar %[[V_36]]#0, %[[V_37:[0-9]+]] : (!fir.ref<!fir.char<1,?>>, index) -> !fir.boxchar<1> |
| ! CHECK: fir.store %[[V_32]] to %[[V_0:[0-9]+]] : !fir.ref<i32> |
| ! CHECK: fir.call @_QPsto_char(%[[V_38]], %[[V_0]]) fastmath<contract> : (!fir.boxchar<1>, !fir.ref<i32>) -> () |
| ! CHECK: fir.result %[[V_36]]#1 : !fir.array<?x?x!fir.char<1,?>> |
| ! CHECK: } |
| ! CHECK: fir.result %[[V_27:[0-9]+]] : !fir.array<?x?x!fir.char<1,?>> |
| ! CHECK: } |
| ! CHECK: fir.array_merge_store %[[V_4]], %[[V_6]] to %arg1 : !fir.array<?x?x!fir.char<1,?>>, !fir.array<?x?x!fir.char<1,?>>, !fir.box<!fir.array<?x?x!fir.char<1,?>>> |
| ! CHECK: return |
| ! CHECK: } |
| |
| subroutine from_char_forall_array(i, c) |
| interface assignment(=) |
| elemental subroutine sfrom_char(a,b) |
| integer, intent(out) :: a |
| character(*),intent(in) :: b |
| end subroutine |
| end interface |
| integer :: i(:, :) |
| character(*) :: c(:, :) |
| forall (j=1:10) i(j, :) = c(j, :) |
| end subroutine |
| |
| subroutine to_char_forall_array(i, c) |
| interface assignment(=) |
| elemental subroutine sto_char(a,b) |
| character(*), intent(out) :: a |
| integer,intent(in) :: b |
| end subroutine |
| end interface |
| integer :: i(:, :) |
| character(*) :: c(:, :) |
| forall (j=1:10) c(j, :) = i(j, :) |
| end subroutine |
| |
| ! TODO: test array user defined assignment inside FORALL. |
| subroutine test_todo(x, y) |
| interface assignment(=) |
| ! User assignment is not elemental, it takes array arguments. |
| pure subroutine assign_array(a,b) |
| logical, intent(out) :: a(:) |
| integer, intent(in) :: b(:) |
| end |
| end interface |
| logical :: x(10, 10) |
| integer :: y(10, 10) |
| ! forall(i=1:10) x(i, :) = y(i, :) |
| end subroutine |
