clang/test/OpenMP/irbuilder_for_iterator.cpp

   1 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --include-generated-funcs --replace-value-regex "__omp_offloading_[0-9a-z]+_[0-9a-z]+" "reduction_size[.].+[.]" "pl_cond[.].+[.|,]" --prefix-filecheck-ir-name _
   2 // RUN: %clang_cc1 -fopenmp-enable-irbuilder -verify -fopenmp -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -o - | FileCheck %s
   3 // expected-no-diagnostics
   4
   5 #ifndef HEADER
   6 #define HEADER
   7
   8 struct MyIterator {
   9   MyIterator(unsigned pos);
  10   MyIterator(const MyIterator &other);
  11   const MyIterator &operator=(const MyIterator &that);
  12   MyIterator &operator++();
  13   int operator-(const MyIterator &that) const;
  14   MyIterator &operator+=(unsigned a);
  15   MyIterator operator+(unsigned a) const;
  16   bool operator==(const MyIterator &that) const;
  17   bool operator!=(const MyIterator &that) const;
  18   unsigned operator*() const;
  19 };
  20
  21 extern "C" void workshareloop_iterator(float *a, float *b, float *c) {
  22 #pragma omp for
  23   for (MyIterator it = MyIterator(7); it != MyIterator(41); ++it) {
  24     unsigned i = *it;
  25     a[i] = b[i] * c[i];
  26   }
  27 }
  28
  29 #endif // HEADER
  30 // CHECK-LABEL: define {{[^@]+}}@workshareloop_iterator
  31 // CHECK-SAME: (ptr noundef [[A:%.*]], ptr noundef [[B:%.*]], ptr noundef [[C:%.*]]) #[[ATTR0:[0-9]+]] {
  32 // CHECK-NEXT:  entry:
  33 // CHECK-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8
  34 // CHECK-NEXT:    [[B_ADDR:%.*]] = alloca ptr, align 8
  35 // CHECK-NEXT:    [[C_ADDR:%.*]] = alloca ptr, align 8
  36 // CHECK-NEXT:    [[IT:%.*]] = alloca [[STRUCT_MYITERATOR:%.*]], align 1
  37 // CHECK-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 8
  38 // CHECK-NEXT:    [[AGG_CAPTURED1:%.*]] = alloca [[STRUCT_ANON_0:%.*]], align 1
  39 // CHECK-NEXT:    [[DOTCOUNT_ADDR:%.*]] = alloca i64, align 8
  40 // CHECK-NEXT:    [[I:%.*]] = alloca i32, align 4
  41 // CHECK-NEXT:    [[P_LASTITER:%.*]] = alloca i32, align 4
  42 // CHECK-NEXT:    [[P_LOWERBOUND:%.*]] = alloca i64, align 8
  43 // CHECK-NEXT:    [[P_UPPERBOUND:%.*]] = alloca i64, align 8
  44 // CHECK-NEXT:    [[P_STRIDE:%.*]] = alloca i64, align 8
  45 // CHECK-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 8
  46 // CHECK-NEXT:    store ptr [[B]], ptr [[B_ADDR]], align 8
  47 // CHECK-NEXT:    store ptr [[C]], ptr [[C_ADDR]], align 8
  48 // CHECK-NEXT:    call void @_ZN10MyIteratorC1Ej(ptr noundef nonnull align 1 dereferenceable(1) [[IT]], i32 noundef 7)
  49 // CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
  50 // CHECK-NEXT:    store ptr [[IT]], ptr [[TMP0]], align 8
  51 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_ANON_0]], ptr [[AGG_CAPTURED1]], i32 0, i32 0
  52 // CHECK-NEXT:    call void @_ZN10MyIteratorC1ERKS_(ptr noundef nonnull align 1 dereferenceable(1) [[TMP1]], ptr noundef nonnull align 1 dereferenceable(1) [[IT]])
  53 // CHECK-NEXT:    call void @__captured_stmt(ptr [[DOTCOUNT_ADDR]], ptr [[AGG_CAPTURED]])
  54 // CHECK-NEXT:    [[DOTCOUNT:%.*]] = load i64, ptr [[DOTCOUNT_ADDR]], align 8
  55 // CHECK-NEXT:    br label [[OMP_LOOP_PREHEADER:%.*]]
  56 // CHECK:       omp_loop.preheader:
  57 // CHECK-NEXT:    store i64 0, ptr [[P_LOWERBOUND]], align 8
  58 // CHECK-NEXT:    [[TMP2:%.*]] = sub i64 [[DOTCOUNT]], 1
  59 // CHECK-NEXT:    store i64 [[TMP2]], ptr [[P_UPPERBOUND]], align 8
  60 // CHECK-NEXT:    store i64 1, ptr [[P_STRIDE]], align 8
  61 // CHECK-NEXT:    [[OMP_GLOBAL_THREAD_NUM:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]])
  62 // CHECK-NEXT:    call void @__kmpc_for_static_init_8u(ptr @[[GLOB1]], i32 [[OMP_GLOBAL_THREAD_NUM]], i32 34, ptr [[P_LASTITER]], ptr [[P_LOWERBOUND]], ptr [[P_UPPERBOUND]], ptr [[P_STRIDE]], i64 1, i64 0)
  63 // CHECK-NEXT:    [[TMP3:%.*]] = load i64, ptr [[P_LOWERBOUND]], align 8
  64 // CHECK-NEXT:    [[TMP4:%.*]] = load i64, ptr [[P_UPPERBOUND]], align 8
  65 // CHECK-NEXT:    [[TMP5:%.*]] = sub i64 [[TMP4]], [[TMP3]]
  66 // CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[TMP5]], 1
  67 // CHECK-NEXT:    br label [[OMP_LOOP_HEADER:%.*]]
  68 // CHECK:       omp_loop.header:
  69 // CHECK-NEXT:    [[OMP_LOOP_IV:%.*]] = phi i64 [ 0, [[OMP_LOOP_PREHEADER]] ], [ [[OMP_LOOP_NEXT:%.*]], [[OMP_LOOP_INC:%.*]] ]
  70 // CHECK-NEXT:    br label [[OMP_LOOP_COND:%.*]]
  71 // CHECK:       omp_loop.cond:
  72 // CHECK-NEXT:    [[OMP_LOOP_CMP:%.*]] = icmp ult i64 [[OMP_LOOP_IV]], [[TMP6]]
  73 // CHECK-NEXT:    br i1 [[OMP_LOOP_CMP]], label [[OMP_LOOP_BODY:%.*]], label [[OMP_LOOP_EXIT:%.*]]
  74 // CHECK:       omp_loop.body:
  75 // CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[OMP_LOOP_IV]], [[TMP3]]
  76 // CHECK-NEXT:    call void @__captured_stmt.1(ptr [[IT]], i64 [[TMP7]], ptr [[AGG_CAPTURED1]])
  77 // CHECK-NEXT:    [[CALL:%.*]] = call noundef i32 @_ZNK10MyIteratordeEv(ptr noundef nonnull align 1 dereferenceable(1) [[IT]])
  78 // CHECK-NEXT:    store i32 [[CALL]], ptr [[I]], align 4
  79 // CHECK-NEXT:    [[TMP8:%.*]] = load ptr, ptr [[B_ADDR]], align 8
  80 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[I]], align 4
  81 // CHECK-NEXT:    [[IDXPROM:%.*]] = zext i32 [[TMP9]] to i64
  82 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[TMP8]], i64 [[IDXPROM]]
  83 // CHECK-NEXT:    [[TMP10:%.*]] = load float, ptr [[ARRAYIDX]], align 4
  84 // CHECK-NEXT:    [[TMP11:%.*]] = load ptr, ptr [[C_ADDR]], align 8
  85 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[I]], align 4
  86 // CHECK-NEXT:    [[IDXPROM2:%.*]] = zext i32 [[TMP12]] to i64
  87 // CHECK-NEXT:    [[ARRAYIDX3:%.*]] = getelementptr inbounds float, ptr [[TMP11]], i64 [[IDXPROM2]]
  88 // CHECK-NEXT:    [[TMP13:%.*]] = load float, ptr [[ARRAYIDX3]], align 4
  89 // CHECK-NEXT:    [[MUL:%.*]] = fmul float [[TMP10]], [[TMP13]]
  90 // CHECK-NEXT:    [[TMP14:%.*]] = load ptr, ptr [[A_ADDR]], align 8
  91 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[I]], align 4
  92 // CHECK-NEXT:    [[IDXPROM4:%.*]] = zext i32 [[TMP15]] to i64
  93 // CHECK-NEXT:    [[ARRAYIDX5:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i64 [[IDXPROM4]]
  94 // CHECK-NEXT:    store float [[MUL]], ptr [[ARRAYIDX5]], align 4
  95 // CHECK-NEXT:    br label [[OMP_LOOP_INC]]
  96 // CHECK:       omp_loop.inc:
  97 // CHECK-NEXT:    [[OMP_LOOP_NEXT]] = add nuw i64 [[OMP_LOOP_IV]], 1
  98 // CHECK-NEXT:    br label [[OMP_LOOP_HEADER]]
  99 // CHECK:       omp_loop.exit:
 100 // CHECK-NEXT:    call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[OMP_GLOBAL_THREAD_NUM]])
 101 // CHECK-NEXT:    [[OMP_GLOBAL_THREAD_NUM6:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
 102 // CHECK-NEXT:    call void @__kmpc_barrier(ptr @[[GLOB2:[0-9]+]], i32 [[OMP_GLOBAL_THREAD_NUM6]])
 103 // CHECK-NEXT:    br label [[OMP_LOOP_AFTER:%.*]]
 104 // CHECK:       omp_loop.after:
 105 // CHECK-NEXT:    ret void
 106 //
 107 //
 108 // CHECK-LABEL: define {{[^@]+}}@__captured_stmt
 109 // CHECK-SAME: (ptr noundef nonnull align 8 dereferenceable(8) [[DISTANCE:%.*]], ptr noalias noundef [[__CONTEXT:%.*]]) #[[ATTR2:[0-9]+]] {
 110 // CHECK-NEXT:  entry:
 111 // CHECK-NEXT:    [[DISTANCE_ADDR:%.*]] = alloca ptr, align 8
 112 // CHECK-NEXT:    [[__CONTEXT_ADDR:%.*]] = alloca ptr, align 8
 113 // CHECK-NEXT:    [[DOTSTART:%.*]] = alloca [[STRUCT_MYITERATOR:%.*]], align 1
 114 // CHECK-NEXT:    [[DOTSTOP:%.*]] = alloca [[STRUCT_MYITERATOR]], align 1
 115 // CHECK-NEXT:    [[DOTSTEP:%.*]] = alloca i64, align 8
 116 // CHECK-NEXT:    store ptr [[DISTANCE]], ptr [[DISTANCE_ADDR]], align 8
 117 // CHECK-NEXT:    store ptr [[__CONTEXT]], ptr [[__CONTEXT_ADDR]], align 8
 118 // CHECK-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[__CONTEXT_ADDR]], align 8
 119 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_ANON:%.*]], ptr [[TMP0]], i32 0, i32 0
 120 // CHECK-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[TMP1]], align 8
 121 // CHECK-NEXT:    call void @_ZN10MyIteratorC1ERKS_(ptr noundef nonnull align 1 dereferenceable(1) [[DOTSTART]], ptr noundef nonnull align 1 dereferenceable(1) [[TMP2]])
 122 // CHECK-NEXT:    call void @_ZN10MyIteratorC1Ej(ptr noundef nonnull align 1 dereferenceable(1) [[DOTSTOP]], i32 noundef 41)
 123 // CHECK-NEXT:    store i64 1, ptr [[DOTSTEP]], align 8
 124 // CHECK-NEXT:    [[TMP3:%.*]] = load i64, ptr [[DOTSTEP]], align 8
 125 // CHECK-NEXT:    [[CMP:%.*]] = icmp ult i64 [[TMP3]], 0
 126 // CHECK-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
 127 // CHECK:       cond.true:
 128 // CHECK-NEXT:    [[CALL:%.*]] = call noundef i32 @_ZNK10MyIteratormiERKS_(ptr noundef nonnull align 1 dereferenceable(1) [[DOTSTART]], ptr noundef nonnull align 1 dereferenceable(1) [[DOTSTOP]])
 129 // CHECK-NEXT:    [[CONV:%.*]] = sext i32 [[CALL]] to i64
 130 // CHECK-NEXT:    [[TMP4:%.*]] = load i64, ptr [[DOTSTEP]], align 8
 131 // CHECK-NEXT:    [[SUB:%.*]] = sub i64 0, [[TMP4]]
 132 // CHECK-NEXT:    [[DIV:%.*]] = udiv i64 [[CONV]], [[SUB]]
 133 // CHECK-NEXT:    br label [[COND_END:%.*]]
 134 // CHECK:       cond.false:
 135 // CHECK-NEXT:    [[CALL1:%.*]] = call noundef i32 @_ZNK10MyIteratormiERKS_(ptr noundef nonnull align 1 dereferenceable(1) [[DOTSTOP]], ptr noundef nonnull align 1 dereferenceable(1) [[DOTSTART]])
 136 // CHECK-NEXT:    [[CONV2:%.*]] = sext i32 [[CALL1]] to i64
 137 // CHECK-NEXT:    [[TMP5:%.*]] = load i64, ptr [[DOTSTEP]], align 8
 138 // CHECK-NEXT:    [[DIV3:%.*]] = udiv i64 [[CONV2]], [[TMP5]]
 139 // CHECK-NEXT:    br label [[COND_END]]
 140 // CHECK:       cond.end:
 141 // CHECK-NEXT:    [[COND:%.*]] = phi i64 [ [[DIV]], [[COND_TRUE]] ], [ [[DIV3]], [[COND_FALSE]] ]
 142 // CHECK-NEXT:    [[TMP6:%.*]] = load ptr, ptr [[DISTANCE_ADDR]], align 8
 143 // CHECK-NEXT:    store i64 [[COND]], ptr [[TMP6]], align 8
 144 // CHECK-NEXT:    ret void
 145 //
 146 //
 147 // CHECK-LABEL: define {{[^@]+}}@__captured_stmt.1
 148 // CHECK-SAME: (ptr noundef nonnull align 1 dereferenceable(1) [[LOOPVAR:%.*]], i64 noundef [[LOGICAL:%.*]], ptr noalias noundef [[__CONTEXT:%.*]]) #[[ATTR2]] {
 149 // CHECK-NEXT:  entry:
 150 // CHECK-NEXT:    [[LOOPVAR_ADDR:%.*]] = alloca ptr, align 8
 151 // CHECK-NEXT:    [[LOGICAL_ADDR:%.*]] = alloca i64, align 8
 152 // CHECK-NEXT:    [[__CONTEXT_ADDR:%.*]] = alloca ptr, align 8
 153 // CHECK-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_MYITERATOR:%.*]], align 1
 154 // CHECK-NEXT:    store ptr [[LOOPVAR]], ptr [[LOOPVAR_ADDR]], align 8
 155 // CHECK-NEXT:    store i64 [[LOGICAL]], ptr [[LOGICAL_ADDR]], align 8
 156 // CHECK-NEXT:    store ptr [[__CONTEXT]], ptr [[__CONTEXT_ADDR]], align 8
 157 // CHECK-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[__CONTEXT_ADDR]], align 8
 158 // CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_ANON_0:%.*]], ptr [[TMP0]], i32 0, i32 0
 159 // CHECK-NEXT:    [[TMP2:%.*]] = load i64, ptr [[LOGICAL_ADDR]], align 8
 160 // CHECK-NEXT:    [[MUL:%.*]] = mul i64 1, [[TMP2]]
 161 // CHECK-NEXT:    [[CONV:%.*]] = trunc i64 [[MUL]] to i32
 162 // CHECK-NEXT:    call void @_ZNK10MyIteratorplEj(ptr sret([[STRUCT_MYITERATOR]]) align 1 [[REF_TMP]], ptr noundef nonnull align 1 dereferenceable(1) [[TMP1]], i32 noundef [[CONV]])
 163 // CHECK-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[LOOPVAR_ADDR]], align 8
 164 // CHECK-NEXT:    [[CALL:%.*]] = call noundef nonnull align 1 dereferenceable(1) ptr @_ZN10MyIteratoraSERKS_(ptr noundef nonnull align 1 dereferenceable(1) [[TMP3]], ptr noundef nonnull align 1 dereferenceable(1) [[REF_TMP]])
 165 // CHECK-NEXT:    ret void
 166 //