// 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 _
// RUN: %clang_cc1 -no-enable-noundef-analysis -verify -triple x86_64-apple-darwin10 -fopenmp -x c++ -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -no-enable-noundef-analysis -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
// RUN: %clang_cc1 -no-enable-noundef-analysis -fopenmp -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK1

// RUN: %clang_cc1 -no-enable-noundef-analysis -verify -triple x86_64-apple-darwin10 -fopenmp-simd -x c++ -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK3
// RUN: %clang_cc1 -no-enable-noundef-analysis -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
// RUN: %clang_cc1 -no-enable-noundef-analysis -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK3
// expected-no-diagnostics
#ifndef HEADER
#define HEADER


struct S {
  int a;
  S() : a(0) {}
  S(const S&) {}
  S& operator=(const S&) {return *this;}
  ~S() {}
  friend S operator+(const S&a, const S&b) {return a;}
};


int main(int argc, char **argv) {
  int a;
  float b;
  S c[5];
  short d[argc];
#pragma omp taskgroup task_reduction(+: a, b, argc)
  {
#pragma omp taskgroup task_reduction(-:c, d)
#pragma omp parallel
#pragma omp taskloop simd in_reduction(+:a) in_reduction(-:d)
    for (int i = 0; i < 5; ++i)
      a += d[a];
  }
  return 0;
}


#endif
// CHECK1-LABEL: define {{[^@]+}}@main
// CHECK1-SAME: (i32 [[ARGC:%.*]], ptr [[ARGV:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[ARGV_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[A:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[B:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[C:%.*]] = alloca [5 x %struct.S], align 16
// CHECK1-NEXT:    [[SAVED_STACK:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTRD_INPUT_:%.*]] = alloca [3 x %struct.kmp_taskred_input_t], align 8
// CHECK1-NEXT:    [[DOTTASK_RED_:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTRD_INPUT_3:%.*]] = alloca [2 x %struct.kmp_taskred_input_t.0], align 8
// CHECK1-NEXT:    [[DOTTASK_RED_6:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]])
// CHECK1-NEXT:    store i32 0, ptr [[RETVAL]], align 4
// CHECK1-NEXT:    store i32 [[ARGC]], ptr [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], ptr [[C]], i32 0, i32 0
// CHECK1-NEXT:    [[ARRAYCTOR_END:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[ARRAY_BEGIN]], i64 5
// CHECK1-NEXT:    br label [[ARRAYCTOR_LOOP:%.*]]
// CHECK1:       arrayctor.loop:
// CHECK1-NEXT:    [[ARRAYCTOR_CUR:%.*]] = phi ptr [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[ARRAYCTOR_NEXT:%.*]], [[ARRAYCTOR_LOOP]] ]
// CHECK1-NEXT:    call void @_ZN1SC1Ev(ptr nonnull align 4 dereferenceable(4) [[ARRAYCTOR_CUR]])
// CHECK1-NEXT:    [[ARRAYCTOR_NEXT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYCTOR_CUR]], i64 1
// CHECK1-NEXT:    [[ARRAYCTOR_DONE:%.*]] = icmp eq ptr [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK1-NEXT:    br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK1:       arrayctor.cont:
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = zext i32 [[TMP1]] to i64
// CHECK1-NEXT:    [[TMP3:%.*]] = call ptr @llvm.stacksave()
// CHECK1-NEXT:    store ptr [[TMP3]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT:    [[VLA:%.*]] = alloca i16, i64 [[TMP2]], align 16
// CHECK1-NEXT:    store i64 [[TMP2]], ptr [[__VLA_EXPR0]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], ptr [[DOTRD_INPUT_]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T:%.*]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
// CHECK1-NEXT:    store ptr [[A]], ptr [[TMP4]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
// CHECK1-NEXT:    store ptr [[A]], ptr [[TMP5]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
// CHECK1-NEXT:    store i64 4, ptr [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
// CHECK1-NEXT:    store ptr @.red_init., ptr [[TMP7]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
// CHECK1-NEXT:    store ptr null, ptr [[TMP8]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
// CHECK1-NEXT:    store ptr @.red_comb., ptr [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
// CHECK1-NEXT:    call void @llvm.memset.p0.i64(ptr align 8 [[TMP10]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_1:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], ptr [[DOTRD_INPUT_]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 0
// CHECK1-NEXT:    store ptr [[B]], ptr [[TMP11]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 1
// CHECK1-NEXT:    store ptr [[B]], ptr [[TMP12]], align 8
// CHECK1-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 2
// CHECK1-NEXT:    store i64 4, ptr [[TMP13]], align 8
// CHECK1-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 3
// CHECK1-NEXT:    store ptr @.red_init..1, ptr [[TMP14]], align 8
// CHECK1-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 4
// CHECK1-NEXT:    store ptr null, ptr [[TMP15]], align 8
// CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 5
// CHECK1-NEXT:    store ptr @.red_comb..2, ptr [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_1]], i32 0, i32 6
// CHECK1-NEXT:    call void @llvm.memset.p0.i64(ptr align 8 [[TMP17]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_2:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], ptr [[DOTRD_INPUT_]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 0
// CHECK1-NEXT:    store ptr [[ARGC_ADDR]], ptr [[TMP18]], align 8
// CHECK1-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 1
// CHECK1-NEXT:    store ptr [[ARGC_ADDR]], ptr [[TMP19]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 2
// CHECK1-NEXT:    store i64 4, ptr [[TMP20]], align 8
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 3
// CHECK1-NEXT:    store ptr @.red_init..3, ptr [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 4
// CHECK1-NEXT:    store ptr null, ptr [[TMP22]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 5
// CHECK1-NEXT:    store ptr @.red_comb..4, ptr [[TMP23]], align 8
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_2]], i32 0, i32 6
// CHECK1-NEXT:    call void @llvm.memset.p0.i64(ptr align 8 [[TMP24]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[TMP25:%.*]] = call ptr @__kmpc_taskred_init(i32 [[TMP0]], i32 3, ptr [[DOTRD_INPUT_]])
// CHECK1-NEXT:    store ptr [[TMP25]], ptr [[DOTTASK_RED_]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_4:%.*]] = getelementptr inbounds [2 x %struct.kmp_taskred_input_t.0], ptr [[DOTRD_INPUT_3]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0:%.*]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 0
// CHECK1-NEXT:    store ptr [[C]], ptr [[TMP26]], align 8
// CHECK1-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 1
// CHECK1-NEXT:    store ptr [[C]], ptr [[TMP27]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK1-NEXT:    store i64 20, ptr [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP29:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK1-NEXT:    store ptr @.red_init..5, ptr [[TMP29]], align 8
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 4
// CHECK1-NEXT:    store ptr @.red_fini., ptr [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP31:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 5
// CHECK1-NEXT:    store ptr @.red_comb..6, ptr [[TMP31]], align 8
// CHECK1-NEXT:    [[TMP32:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 6
// CHECK1-NEXT:    call void @llvm.memset.p0.i64(ptr align 8 [[TMP32]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_5:%.*]] = getelementptr inbounds [2 x %struct.kmp_taskred_input_t.0], ptr [[DOTRD_INPUT_3]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP33:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 0
// CHECK1-NEXT:    store ptr [[VLA]], ptr [[TMP33]], align 8
// CHECK1-NEXT:    [[TMP34:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT:    store ptr [[VLA]], ptr [[TMP34]], align 8
// CHECK1-NEXT:    [[TMP35:%.*]] = mul nuw i64 [[TMP2]], 2
// CHECK1-NEXT:    [[TMP36:%.*]] = udiv exact i64 [[TMP35]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
// CHECK1-NEXT:    [[TMP37:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT:    store i64 [[TMP35]], ptr [[TMP37]], align 8
// CHECK1-NEXT:    [[TMP38:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
// CHECK1-NEXT:    store ptr @.red_init..7, ptr [[TMP38]], align 8
// CHECK1-NEXT:    [[TMP39:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 4
// CHECK1-NEXT:    store ptr null, ptr [[TMP39]], align 8
// CHECK1-NEXT:    [[TMP40:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 5
// CHECK1-NEXT:    store ptr @.red_comb..8, ptr [[TMP40]], align 8
// CHECK1-NEXT:    [[TMP41:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 6
// CHECK1-NEXT:    store i32 1, ptr [[TMP41]], align 8
// CHECK1-NEXT:    [[TMP42:%.*]] = call ptr @__kmpc_taskred_init(i32 [[TMP0]], i32 2, ptr [[DOTRD_INPUT_3]])
// CHECK1-NEXT:    store ptr [[TMP42]], ptr [[DOTTASK_RED_6]], align 8
// CHECK1-NEXT:    call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 5, ptr @main.omp_outlined, ptr [[A]], i64 [[TMP2]], ptr [[VLA]], ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED_6]])
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(ptr @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    store i32 0, ptr [[RETVAL]], align 4
// CHECK1-NEXT:    [[TMP43:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT:    call void @llvm.stackrestore(ptr [[TMP43]])
// CHECK1-NEXT:    [[ARRAY_BEGIN7:%.*]] = getelementptr inbounds [5 x %struct.S], ptr [[C]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP44:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAY_BEGIN7]], i64 5
// CHECK1-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1:       arraydestroy.body:
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP44]], [[ARRAYCTOR_CONT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT:    call void @_ZN1SD1Ev(ptr nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3:[0-9]+]]
// CHECK1-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN7]]
// CHECK1-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE8:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK1:       arraydestroy.done8:
// CHECK1-NEXT:    [[TMP45:%.*]] = load i32, ptr [[RETVAL]], align 4
// CHECK1-NEXT:    ret i32 [[TMP45]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC1Ev
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1:[0-9]+]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SC2Ev(ptr nonnull align 4 dereferenceable(4) [[THIS1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init.
// CHECK1-SAME: (ptr noalias [[TMP0:%.*]], ptr noalias [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store i32 0, ptr [[TMP2]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb.
// CHECK1-SAME: (ptr [[TMP0:%.*]], ptr [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, ptr [[TMP2]], align 4
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32, ptr [[TMP3]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP4]], [[TMP5]]
// CHECK1-NEXT:    store i32 [[ADD]], ptr [[TMP2]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..1
// CHECK1-SAME: (ptr noalias [[TMP0:%.*]], ptr noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store float 0.000000e+00, ptr [[TMP2]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..2
// CHECK1-SAME: (ptr [[TMP0:%.*]], ptr [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load float, ptr [[TMP2]], align 4
// CHECK1-NEXT:    [[TMP5:%.*]] = load float, ptr [[TMP3]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = fadd float [[TMP4]], [[TMP5]]
// CHECK1-NEXT:    store float [[ADD]], ptr [[TMP2]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..3
// CHECK1-SAME: (ptr noalias [[TMP0:%.*]], ptr noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store i32 0, ptr [[TMP2]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..4
// CHECK1-SAME: (ptr [[TMP0:%.*]], ptr [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, ptr [[TMP2]], align 4
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32, ptr [[TMP3]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP4]], [[TMP5]]
// CHECK1-NEXT:    store i32 [[ADD]], ptr [[TMP2]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..5
// CHECK1-SAME: (ptr noalias [[TMP0:%.*]], ptr noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], ptr [[TMP2]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP3:%.*]] = getelementptr [[STRUCT_S:%.*]], ptr [[ARRAY_BEGIN]], i64 5
// CHECK1-NEXT:    [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[ARRAY_BEGIN]], [[TMP3]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1:       omp.arrayinit.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT:    call void @_ZN1SC1Ev(ptr nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]])
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP3]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1:       omp.arrayinit.done:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_fini.
// CHECK1-SAME: (ptr [[TMP0:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], ptr [[TMP1]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP2:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[ARRAY_BEGIN]], i64 5
// CHECK1-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1:       arraydestroy.body:
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP2]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT:    call void @_ZN1SD1Ev(ptr nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]]
// CHECK1-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN]]
// CHECK1-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE1:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK1:       arraydestroy.done1:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SD1Ev
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SD2Ev(ptr nonnull align 4 dereferenceable(4) [[THIS1]]) #[[ATTR3]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..6
// CHECK1-SAME: (ptr [[TMP0:%.*]], ptr [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr [[STRUCT_S]], ptr [[TMP2]], i64 5
// CHECK1-NEXT:    [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[TMP2]], [[TMP4]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE2:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1:       omp.arraycpy.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[TMP3]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[TMP2]], [[ENTRY]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    call void @_ZplRK1SS1_(ptr sret([[STRUCT_S]]) align 4 [[REF_TMP]], ptr nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]], ptr nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_SRCELEMENTPAST]])
// CHECK1-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) ptr @_ZN1SaSERKS_(ptr nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]], ptr nonnull align 4 dereferenceable(4) [[REF_TMP]])
// CHECK1-NEXT:    call void @_ZN1SD1Ev(ptr nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR3]]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP4]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYCPY_DONE2]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1:       omp.arraycpy.done2:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZplRK1SS1_
// CHECK1-SAME: (ptr noalias sret([[STRUCT_S:%.*]]) align 4 [[AGG_RESULT:%.*]], ptr nonnull align 4 dereferenceable(4) [[A:%.*]], ptr nonnull align 4 dereferenceable(4) [[B:%.*]]) #[[ATTR7:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[RESULT_PTR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[B_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 8
// CHECK1-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SC1ERKS_(ptr nonnull align 4 dereferenceable(4) [[AGG_RESULT]], ptr nonnull align 4 dereferenceable(4) [[TMP0]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SaSERKS_
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]], ptr nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR7]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret ptr [[THIS1]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..7
// CHECK1-SAME: (ptr noalias [[TMP0:%.*]], ptr noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP2]], ptr @{{reduction_size[.].+[.]}})
// CHECK1-NEXT:    [[TMP5:%.*]] = load i64, ptr [[TMP4]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr i16, ptr [[TMP3]], i64 [[TMP5]]
// CHECK1-NEXT:    [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[TMP3]], [[TMP6]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1:       omp.arrayinit.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[TMP3]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT:    store i16 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP6]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1:       omp.arrayinit.done:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..8
// CHECK1-SAME: (ptr [[TMP0:%.*]], ptr [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP2]], ptr @{{reduction_size[.].+[.]}})
// CHECK1-NEXT:    [[TMP4:%.*]] = load i64, ptr [[TMP3]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr i16, ptr [[TMP5]], i64 [[TMP4]]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[TMP5]], [[TMP7]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE4:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1:       omp.arraycpy.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[TMP6]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[TMP5]], [[ENTRY]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    [[TMP8:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT:    [[CONV:%.*]] = sext i16 [[TMP8]] to i32
// CHECK1-NEXT:    [[TMP9:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
// CHECK1-NEXT:    [[CONV2:%.*]] = sext i16 [[TMP9]] to i32
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], [[CONV2]]
// CHECK1-NEXT:    [[CONV3:%.*]] = trunc i32 [[ADD]] to i16
// CHECK1-NEXT:    store i16 [[CONV3]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP7]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYCPY_DONE4]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1:       omp.arraycpy.done4:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@main.omp_outlined
// CHECK1-SAME: (ptr noalias [[DOTGLOBAL_TID_:%.*]], ptr noalias [[DOTBOUND_TID_:%.*]], ptr nonnull align 4 dereferenceable(4) [[A:%.*]], i64 [[VLA:%.*]], ptr nonnull align 2 dereferenceable(2) [[D:%.*]], ptr nonnull align 8 dereferenceable(8) [[DOTTASK_RED_:%.*]], ptr nonnull align 8 dereferenceable(8) [[DOTTASK_RED_1:%.*]]) #[[ATTR8:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[D_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTTASK_RED__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTTASK_RED__ADDR2:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 8
// CHECK1-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[D]], ptr [[D_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT:    store ptr [[DOTTASK_RED_1]], ptr [[DOTTASK_RED__ADDR2]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[TMP5]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
// CHECK1-NEXT:    store i64 [[TMP1]], ptr [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
// CHECK1-NEXT:    store ptr [[TMP2]], ptr [[TMP7]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 3
// CHECK1-NEXT:    store ptr [[TMP3]], ptr [[TMP8]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 4
// CHECK1-NEXT:    store ptr [[TMP4]], ptr [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = load i32, ptr [[TMP10]], align 4
// CHECK1-NEXT:    call void @__kmpc_taskgroup(ptr @[[GLOB1]], i32 [[TMP11]])
// CHECK1-NEXT:    [[TMP12:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP11]], i32 1, i64 96, i64 40, ptr @.omp_task_entry.)
// CHECK1-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP12]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP13]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
// CHECK1-NEXT:    call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP15]], ptr align 8 [[AGG_CAPTURED]], i64 40, i1 false)
// CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP12]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP16]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP18:%.*]] = load ptr, ptr [[TMP3]], align 8
// CHECK1-NEXT:    store ptr [[TMP18]], ptr [[TMP17]], align 8
// CHECK1-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T]], ptr [[TMP16]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP20:%.*]] = load ptr, ptr [[TMP4]], align 8
// CHECK1-NEXT:    store ptr [[TMP20]], ptr [[TMP19]], align 8
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP13]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, ptr [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP13]], i32 0, i32 6
// CHECK1-NEXT:    store i64 4, ptr [[TMP22]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP13]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, ptr [[TMP23]], align 8
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP13]], i32 0, i32 9
// CHECK1-NEXT:    call void @llvm.memset.p0.i64(ptr align 8 [[TMP24]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP25:%.*]] = load i64, ptr [[TMP23]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskloop(ptr @[[GLOB1]], i32 [[TMP11]], ptr [[TMP12]], i32 1, ptr [[TMP21]], ptr [[TMP22]], i64 [[TMP25]], i32 1, i32 0, i64 0, ptr null)
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(ptr @[[GLOB1]], i32 [[TMP11]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_privates_map.
// CHECK1-SAME: (ptr noalias [[TMP0:%.*]], ptr noalias [[TMP1:%.*]], ptr noalias [[TMP2:%.*]]) #[[ATTR9:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR2:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    store ptr [[TMP2]], ptr [[DOTADDR2]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP3]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    store ptr [[TMP4]], ptr [[TMP5]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T]], ptr [[TMP3]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP7:%.*]] = load ptr, ptr [[DOTADDR2]], align 8
// CHECK1-NEXT:    store ptr [[TMP6]], ptr [[TMP7]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry.
// CHECK1-SAME: (i32 [[TMP0:%.*]], ptr noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTPART_ID__ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTPRIVATES__ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTCOPY_FN__ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTTASK_T__ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTLB__ADDR_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTUB__ADDR_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTST__ADDR_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTLITER__ADDR_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTREDUCTIONS__ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[__CONTEXT_ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTFIRSTPRIV_PTR_ADDR_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTFIRSTPRIV_PTR_ADDR1_I:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store i32 [[TMP0]], ptr [[DOTADDR]], align 4
// CHECK1-NEXT:    store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, ptr [[DOTADDR]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP3]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP4]], i32 0, i32 2
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP10:%.*]] = load i64, ptr [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP12:%.*]] = load i64, ptr [[TMP11]], align 8
// CHECK1-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 7
// CHECK1-NEXT:    [[TMP14:%.*]] = load i64, ptr [[TMP13]], align 8
// CHECK1-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 8
// CHECK1-NEXT:    [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 8
// CHECK1-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP18:%.*]] = load ptr, ptr [[TMP17]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META3:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !14
// CHECK1-NEXT:    store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias !14
// CHECK1-NEXT:    store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_ANON:%.*]], ptr [[TMP19]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP21:%.*]] = load i64, ptr [[TMP20]], align 8
// CHECK1-NEXT:    [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
// CHECK1-NEXT:    [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT:    [[TMP27:%.*]] = load ptr, ptr [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !14
// CHECK1-NEXT:    [[TMP29:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP27]], ptr [[TMP26]])
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP19]], i32 0, i32 2
// CHECK1-NEXT:    [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP32:%.*]] = mul nuw i64 [[TMP21]], 2
// CHECK1-NEXT:    [[TMP33:%.*]] = udiv exact i64 [[TMP32]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
// CHECK1-NEXT:    [[TMP34:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
// CHECK1-NEXT:    store i64 [[TMP33]], ptr [[TMP34]], align 8
// CHECK1-NEXT:    [[TMP35:%.*]] = load ptr, ptr [[TMP25]], align 8
// CHECK1-NEXT:    [[TMP36:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP35]], ptr [[TMP31]])
// CHECK1-NEXT:    [[TMP37:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP37]] to i32
// CHECK1-NEXT:    store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP38:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias !14, !llvm.access.group [[ACC_GRP15:![0-9]+]]
// CHECK1-NEXT:    [[CONV2_I:%.*]] = sext i32 [[TMP38]] to i64
// CHECK1-NEXT:    [[TMP39:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias !14, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP39]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__EXIT:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP40:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias !14, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    store i32 [[TMP40]], ptr [[I_I]], align 4, !noalias !14, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[TMP41:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[IDXPROM_I:%.*]] = sext i32 [[TMP41]] to i64
// CHECK1-NEXT:    [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP36]], i64 [[IDXPROM_I]]
// CHECK1-NEXT:    [[TMP42:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[CONV3_I:%.*]] = sext i16 [[TMP42]] to i32
// CHECK1-NEXT:    [[TMP43:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[ADD4_I:%.*]] = add nsw i32 [[TMP43]], [[CONV3_I]]
// CHECK1-NEXT:    store i32 [[ADD4_I]], ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[TMP44:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias !14, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    [[ADD5_I:%.*]] = add nsw i32 [[TMP44]], 1
// CHECK1-NEXT:    store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias !14, !llvm.access.group [[ACC_GRP15]]
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]], !llvm.loop [[LOOP16:![0-9]+]]
// CHECK1:       .omp_outlined..exit:
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC2Ev
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[A:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[THIS1]], i32 0, i32 0
// CHECK1-NEXT:    store i32 0, ptr [[A]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SD2Ev
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC1ERKS_
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]], ptr nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SC2ERKS_(ptr nonnull align 4 dereferenceable(4) [[THIS1]], ptr nonnull align 4 dereferenceable(4) [[TMP1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC2ERKS_
// CHECK1-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]], ptr nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@main
// CHECK3-SAME: (i32 [[ARGC:%.*]], ptr [[ARGV:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[ARGV_ADDR:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    [[A:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[B:%.*]] = alloca float, align 4
// CHECK3-NEXT:    [[C:%.*]] = alloca [5 x %struct.S], align 16
// CHECK3-NEXT:    [[SAVED_STACK:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8
// CHECK3-NEXT:    [[DOTTASK_RED_:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    [[DOTTASK_RED_1:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTOMP_LB:%.*]] = alloca i64, align 8
// CHECK3-NEXT:    [[DOTOMP_UB:%.*]] = alloca i64, align 8
// CHECK3-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32 0, ptr [[RETVAL]], align 4
// CHECK3-NEXT:    store i32 [[ARGC]], ptr [[ARGC_ADDR]], align 4
// CHECK3-NEXT:    store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
// CHECK3-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], ptr [[C]], i32 0, i32 0
// CHECK3-NEXT:    [[ARRAYCTOR_END:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[ARRAY_BEGIN]], i64 5
// CHECK3-NEXT:    br label [[ARRAYCTOR_LOOP:%.*]]
// CHECK3:       arrayctor.loop:
// CHECK3-NEXT:    [[ARRAYCTOR_CUR:%.*]] = phi ptr [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[ARRAYCTOR_NEXT:%.*]], [[ARRAYCTOR_LOOP]] ]
// CHECK3-NEXT:    call void @_ZN1SC1Ev(ptr nonnull align 4 dereferenceable(4) [[ARRAYCTOR_CUR]])
// CHECK3-NEXT:    [[ARRAYCTOR_NEXT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYCTOR_CUR]], i64 1
// CHECK3-NEXT:    [[ARRAYCTOR_DONE:%.*]] = icmp eq ptr [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK3-NEXT:    br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK3:       arrayctor.cont:
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK3-NEXT:    [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
// CHECK3-NEXT:    [[TMP2:%.*]] = call ptr @llvm.stacksave()
// CHECK3-NEXT:    store ptr [[TMP2]], ptr [[SAVED_STACK]], align 8
// CHECK3-NEXT:    [[VLA:%.*]] = alloca i16, i64 [[TMP1]], align 16
// CHECK3-NEXT:    store i64 [[TMP1]], ptr [[__VLA_EXPR0]], align 8
// CHECK3-NEXT:    store i64 0, ptr [[DOTOMP_LB]], align 8
// CHECK3-NEXT:    store i64 4, ptr [[DOTOMP_UB]], align 8
// CHECK3-NEXT:    [[TMP3:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
// CHECK3-NEXT:    [[CONV:%.*]] = trunc i64 [[TMP3]] to i32
// CHECK3-NEXT:    store i32 [[CONV]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3:       omp.inner.for.cond:
// CHECK3-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2:![0-9]+]]
// CHECK3-NEXT:    [[CONV2:%.*]] = sext i32 [[TMP4]] to i64
// CHECK3-NEXT:    [[TMP5:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[CMP:%.*]] = icmp ule i64 [[CONV2]], [[TMP5]]
// CHECK3-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK3:       omp.inner.for.body:
// CHECK3-NEXT:    [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP6]], 1
// CHECK3-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK3-NEXT:    store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[TMP7:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP7]] to i64
// CHECK3-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i16, ptr [[VLA]], i64 [[IDXPROM]]
// CHECK3-NEXT:    [[TMP8:%.*]] = load i16, ptr [[ARRAYIDX]], align 2, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[CONV3:%.*]] = sext i16 [[TMP8]] to i32
// CHECK3-NEXT:    [[TMP9:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[ADD4:%.*]] = add nsw i32 [[TMP9]], [[CONV3]]
// CHECK3-NEXT:    store i32 [[ADD4]], ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK3:       omp.body.continue:
// CHECK3-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK3:       omp.inner.for.inc:
// CHECK3-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    [[ADD5:%.*]] = add nsw i32 [[TMP10]], 1
// CHECK3-NEXT:    store i32 [[ADD5]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
// CHECK3-NEXT:    br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
// CHECK3:       omp.inner.for.end:
// CHECK3-NEXT:    store i32 5, ptr [[I]], align 4
// CHECK3-NEXT:    store i32 0, ptr [[RETVAL]], align 4
// CHECK3-NEXT:    [[TMP11:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
// CHECK3-NEXT:    call void @llvm.stackrestore(ptr [[TMP11]])
// CHECK3-NEXT:    [[ARRAY_BEGIN6:%.*]] = getelementptr inbounds [5 x %struct.S], ptr [[C]], i32 0, i32 0
// CHECK3-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAY_BEGIN6]], i64 5
// CHECK3-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK3:       arraydestroy.body:
// CHECK3-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP12]], [[OMP_INNER_FOR_END]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK3-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK3-NEXT:    call void @_ZN1SD1Ev(ptr nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3:[0-9]+]]
// CHECK3-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN6]]
// CHECK3-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE7:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK3:       arraydestroy.done7:
// CHECK3-NEXT:    [[TMP13:%.*]] = load i32, ptr [[RETVAL]], align 4
// CHECK3-NEXT:    ret i32 [[TMP13]]
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SC1Ev
// CHECK3-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1:[0-9]+]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    call void @_ZN1SC2Ev(ptr nonnull align 4 dereferenceable(4) [[THIS1]])
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SD1Ev
// CHECK3-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    call void @_ZN1SD2Ev(ptr nonnull align 4 dereferenceable(4) [[THIS1]]) #[[ATTR3]]
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SC2Ev
// CHECK3-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[A:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[THIS1]], i32 0, i32 0
// CHECK3-NEXT:    store i32 0, ptr [[A]], align 4
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SD2Ev
// CHECK3-SAME: (ptr nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca ptr, align 8
// CHECK3-NEXT:    store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
// CHECK3-NEXT:    ret void
//