// 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 -verify -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap %s --check-prefix=CHECK1
// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple x86_64-apple-darwin10 -emit-pch -o %t %s
// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap %s --check-prefix=CHECK2
// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap %s --check-prefix=CHECK3
// RUN: %clang_cc1 -verify -fopenmp -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap %s --check-prefix=CHECK4

// RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple x86_64-apple-darwin10 -emit-pch -o %t %s
// RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// expected-no-diagnostics
#ifndef HEADER
#define HEADER

volatile double g;

template <class T>
struct S {
  T f;
  S(T a) : f(a + g) {}
  S() : f(g) {}
  operator T() { return T(); }
  S &operator&(const S &) { return *this; }
  ~S() {}
};


template <typename T>
T tmain() {
  T t;
  S<T> test;
  T t_var = T(), t_var1;
  T vec[] = {1, 2};
  S<T> s_arr[] = {1, 2};
  S<T> var(3), var1;
#pragma omp parallel
#pragma omp sections reduction(+:t_var) reduction(&:var) reduction(&& : var1) reduction(min: t_var1) nowait
  {
    vec[0] = t_var;
#pragma omp section
    s_arr[0] = var;
  }
  return T();
}

int main() {
#ifdef LAMBDA
  [&]() {
#pragma omp parallel
#pragma omp sections reduction(+:g)
    {

    // Reduction list for runtime.

    g = 1;

#pragma omp section
    [&]() {
      g = 2;
    }();
  }
  }();
  return 0;
#elif defined(BLOCKS)
  ^{
#pragma omp parallel
#pragma omp sections reduction(-:g)
    {

    // Reduction list for runtime.

    g = 1;

#pragma omp section
    ^{
      g = 2;
    }();
  }
  }();
  return 0;
#else
  S<float> test;
  float t_var = 0, t_var1;
  int vec[] = {1, 2};
  S<float> s_arr[] = {1, 2};
  S<float> var(3), var1;
#pragma omp parallel
#pragma omp sections reduction(+:t_var) reduction(&:var) reduction(&& : var1) reduction(min: t_var1)
  {
    {
    vec[0] = t_var;
    s_arr[0] = var;
    vec[1] = t_var1;
    s_arr[1] = var1;
    }
  }
  return tmain<int>();
#endif
}


// Reduction list for runtime.


// For + reduction operation initial value of private variable is 0.

// For & reduction operation initial value of private variable is ones in all bits.

// For && reduction operation initial value of private variable is 1.0.

// For min reduction operation initial value of private variable is largest repesentable value.

// Skip checks for internal operations.

// void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};


// res = __kmpc_reduce_nowait(<loc>, <gtid>, <n>, sizeof(RedList), RedList, reduce_func, &<lock>);


// switch(res)

// case 1:
// t_var += t_var_reduction;

// var = var.operator &(var_reduction);

// var1 = var1.operator &&(var1_reduction);

// t_var1 = min(t_var1, t_var1_reduction);

// __kmpc_end_reduce_nowait(<loc>, <gtid>, &<lock>);

// break;

// case 2:
// t_var += t_var_reduction;

// var = var.operator &(var_reduction);

// var1 = var1.operator &&(var1_reduction);

// t_var1 = min(t_var1, t_var1_reduction);

// break;

// void reduce_func(void *lhs[<n>], void *rhs[<n>]) {
//  *(Type0*)lhs[0] = ReductionOperation0(*(Type0*)lhs[0], *(Type0*)rhs[0]);
//  ...
//  *(Type<n>-1*)lhs[<n>-1] = ReductionOperation<n>-1(*(Type<n>-1*)lhs[<n>-1],
//  *(Type<n>-1*)rhs[<n>-1]);
// }
// t_var_lhs = (i{{[0-9]+}}*)lhs[0];
// t_var_rhs = (i{{[0-9]+}}*)rhs[0];

// var_lhs = (S<i{{[0-9]+}}>*)lhs[1];
// var_rhs = (S<i{{[0-9]+}}>*)rhs[1];

// var1_lhs = (S<i{{[0-9]+}}>*)lhs[2];
// var1_rhs = (S<i{{[0-9]+}}>*)rhs[2];

// t_var1_lhs = (i{{[0-9]+}}*)lhs[3];
// t_var1_rhs = (i{{[0-9]+}}*)rhs[3];

// t_var_lhs += t_var_rhs;

// var_lhs = var_lhs.operator &(var_rhs);

// var1_lhs = var1_lhs.operator &&(var1_rhs);

// t_var1_lhs = min(t_var1_lhs, t_var1_rhs);

#endif
// CHECK1-LABEL: define {{[^@]+}}@main
// CHECK1-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TEST:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK1-NEXT:    [[T_VAR:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[T_VAR1:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[VEC:%.*]] = alloca [2 x i32], align 4
// CHECK1-NEXT:    [[S_ARR:%.*]] = alloca [2 x %struct.S], align 4
// CHECK1-NEXT:    [[VAR:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK1-NEXT:    [[VAR1:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK1-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[TEST]])
// CHECK1-NEXT:    store float 0.000000e+00, float* [[T_VAR]], align 4
// CHECK1-NEXT:    [[TMP0:%.*]] = bitcast [2 x i32]* [[VEC]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP0]], i8* align 4 bitcast ([2 x i32]* @__const.main.vec to i8*), i64 8, i1 false)
// CHECK1-NEXT:    [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[S_ARR]], i64 0, i64 0
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float 1.000000e+00)
// CHECK1-NEXT:    [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYINIT_BEGIN]], i64 1
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float 2.000000e+00)
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[VAR]], float 3.000000e+00)
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR1]])
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB3:[0-9]+]], i32 6, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float*, %struct.S*, %struct.S*, float*, [2 x i32]*, [2 x %struct.S]*)* @.omp_outlined. to void (i32*, i32*, ...)*), float* [[T_VAR]], %struct.S* [[VAR]], %struct.S* [[VAR1]], float* [[T_VAR1]], [2 x i32]* [[VEC]], [2 x %struct.S]* [[S_ARR]])
// CHECK1-NEXT:    [[CALL:%.*]] = call i32 @_Z5tmainIiET_v()
// CHECK1-NEXT:    store i32 [[CALL]], i32* [[RETVAL]], align 4
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR1]]) #[[ATTR4:[0-9]+]]
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR]]) #[[ATTR4]]
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[S_ARR]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAY_BEGIN]], i64 2
// CHECK1-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1:       arraydestroy.body:
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP1]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR4]]
// CHECK1-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN]]
// CHECK1-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE1:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK1:       arraydestroy.done1:
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[TEST]]) #[[ATTR4]]
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    ret i32 [[TMP2]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfEC1Ev
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1:[0-9]+]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SIfEC2Ev(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfEC1Ef
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], float [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca float, align 4
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store float [[A]], float* [[A_ADDR]], align 4
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load float, float* [[A_ADDR]], align 4
// CHECK1-NEXT:    call void @_ZN1SIfEC2Ef(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]], float [[TMP0]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], float* nonnull align 4 dereferenceable(4) [[T_VAR:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR1:%.*]], float* nonnull align 4 dereferenceable(4) [[T_VAR1:%.*]], [2 x i32]* nonnull align 4 dereferenceable(8) [[VEC:%.*]], [2 x %struct.S]* nonnull align 4 dereferenceable(8) [[S_ARR:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[T_VAR_ADDR:%.*]] = alloca float*, align 8
// CHECK1-NEXT:    [[VAR_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[VAR1_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[T_VAR1_ADDR:%.*]] = alloca float*, align 8
// CHECK1-NEXT:    [[VEC_ADDR:%.*]] = alloca [2 x i32]*, align 8
// CHECK1-NEXT:    [[S_ARR_ADDR:%.*]] = alloca [2 x %struct.S]*, align 8
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_LB_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_UB_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_ST_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_IL_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_IV_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[T_VAR2:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[VAR3:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK1-NEXT:    [[VAR14:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK1-NEXT:    [[T_VAR15:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [4 x i8*], align 8
// CHECK1-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK1-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[TMP:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[REF_TMP17:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK1-NEXT:    [[ATOMIC_TEMP27:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[_TMP28:%.*]] = alloca float, align 4
// CHECK1-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT:    store float* [[T_VAR]], float** [[T_VAR_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S* [[VAR]], %struct.S** [[VAR_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S* [[VAR1]], %struct.S** [[VAR1_ADDR]], align 8
// CHECK1-NEXT:    store float* [[T_VAR1]], float** [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT:    store [2 x i32]* [[VEC]], [2 x i32]** [[VEC_ADDR]], align 8
// CHECK1-NEXT:    store [2 x %struct.S]* [[S_ARR]], [2 x %struct.S]** [[S_ARR_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load float*, float** [[T_VAR_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load %struct.S*, %struct.S** [[VAR_ADDR]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load %struct.S*, %struct.S** [[VAR1_ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load float*, float** [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load [2 x i32]*, [2 x i32]** [[VEC_ADDR]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = load [2 x %struct.S]*, [2 x %struct.S]** [[S_ARR_ADDR]], align 8
// CHECK1-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_ST_]], align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_IL_]], align 4
// CHECK1-NEXT:    store float 0.000000e+00, float* [[T_VAR2]], align 4
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK1-NEXT:    store float 0x47EFFFFFE0000000, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[TMP6:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4
// CHECK1-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB1:[0-9]+]], i32 [[TMP7]], i32 34, i32* [[DOTOMP_SECTIONS_IL_]], i32* [[DOTOMP_SECTIONS_LB_]], i32* [[DOTOMP_SECTIONS_UB_]], i32* [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
// CHECK1-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    [[TMP9:%.*]] = icmp slt i32 [[TMP8]], 0
// CHECK1-NEXT:    [[TMP10:%.*]] = select i1 [[TMP9]], i32 [[TMP8]], i32 0
// CHECK1-NEXT:    store i32 [[TMP10]], i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    [[TMP11:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK1-NEXT:    store i32 [[TMP11]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1:       omp.inner.for.cond:
// CHECK1-NEXT:    [[TMP12:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    [[TMP13:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    [[CMP:%.*]] = icmp sle i32 [[TMP12]], [[TMP13]]
// CHECK1-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1:       omp.inner.for.body:
// CHECK1-NEXT:    [[TMP14:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    switch i32 [[TMP14]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK1-NEXT:    i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK1-NEXT:    ]
// CHECK1:       .omp.sections.case:
// CHECK1-NEXT:    [[TMP15:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[CONV:%.*]] = fptosi float [[TMP15]] to i32
// CHECK1-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[TMP4]], i64 0, i64 0
// CHECK1-NEXT:    store i32 [[CONV]], i32* [[ARRAYIDX]], align 4
// CHECK1-NEXT:    [[ARRAYIDX6:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[TMP5]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP16:%.*]] = bitcast %struct.S* [[ARRAYIDX6]] to i8*
// CHECK1-NEXT:    [[TMP17:%.*]] = bitcast %struct.S* [[VAR3]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP16]], i8* align 4 [[TMP17]], i64 4, i1 false)
// CHECK1-NEXT:    [[TMP18:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[CONV7:%.*]] = fptosi float [[TMP18]] to i32
// CHECK1-NEXT:    [[ARRAYIDX8:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[TMP4]], i64 0, i64 1
// CHECK1-NEXT:    store i32 [[CONV7]], i32* [[ARRAYIDX8]], align 4
// CHECK1-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[TMP5]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP19:%.*]] = bitcast %struct.S* [[ARRAYIDX9]] to i8*
// CHECK1-NEXT:    [[TMP20:%.*]] = bitcast %struct.S* [[VAR14]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP19]], i8* align 4 [[TMP20]], i64 4, i1 false)
// CHECK1-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK1:       .omp.sections.exit:
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1:       omp.inner.for.inc:
// CHECK1-NEXT:    [[TMP21:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP21]], 1
// CHECK1-NEXT:    store i32 [[INC]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK1:       omp.inner.for.end:
// CHECK1-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[TMP7]])
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP23:%.*]] = bitcast float* [[T_VAR2]] to i8*
// CHECK1-NEXT:    store i8* [[TMP23]], i8** [[TMP22]], align 8
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP25:%.*]] = bitcast %struct.S* [[VAR3]] to i8*
// CHECK1-NEXT:    store i8* [[TMP25]], i8** [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP27:%.*]] = bitcast %struct.S* [[VAR14]] to i8*
// CHECK1-NEXT:    store i8* [[TMP27]], i8** [[TMP26]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
// CHECK1-NEXT:    [[TMP29:%.*]] = bitcast float* [[T_VAR15]] to i8*
// CHECK1-NEXT:    store i8* [[TMP29]], i8** [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP30:%.*]] = bitcast [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT:    [[TMP31:%.*]] = call i32 @__kmpc_reduce(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP7]], i32 4, i64 32, i8* [[TMP30]], void (i8*, i8*)* @.omp.reduction.reduction_func, [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT:    switch i32 [[TMP31]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK1-NEXT:    i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK1-NEXT:    i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK1-NEXT:    ]
// CHECK1:       .omp.reduction.case1:
// CHECK1-NEXT:    [[TMP32:%.*]] = load float, float* [[TMP0]], align 4
// CHECK1-NEXT:    [[TMP33:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = fadd float [[TMP32]], [[TMP33]]
// CHECK1-NEXT:    store float [[ADD]], float* [[TMP0]], align 4
// CHECK1-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SIfEanERKS0_(%struct.S* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT:    [[TMP34:%.*]] = bitcast %struct.S* [[TMP1]] to i8*
// CHECK1-NEXT:    [[TMP35:%.*]] = bitcast %struct.S* [[CALL]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP34]], i8* align 4 [[TMP35]], i64 4, i1 false)
// CHECK1-NEXT:    [[CALL10:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK1-NEXT:    [[TOBOOL:%.*]] = fcmp une float [[CALL10]], 0.000000e+00
// CHECK1-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK1:       land.rhs:
// CHECK1-NEXT:    [[CALL11:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK1-NEXT:    [[TOBOOL12:%.*]] = fcmp une float [[CALL11]], 0.000000e+00
// CHECK1-NEXT:    br label [[LAND_END]]
// CHECK1:       land.end:
// CHECK1-NEXT:    [[TMP36:%.*]] = phi i1 [ false, [[DOTOMP_REDUCTION_CASE1]] ], [ [[TOBOOL12]], [[LAND_RHS]] ]
// CHECK1-NEXT:    [[CONV13:%.*]] = uitofp i1 [[TMP36]] to float
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]], float [[CONV13]])
// CHECK1-NEXT:    [[TMP37:%.*]] = bitcast %struct.S* [[TMP2]] to i8*
// CHECK1-NEXT:    [[TMP38:%.*]] = bitcast %struct.S* [[REF_TMP]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP37]], i8* align 4 [[TMP38]], i64 4, i1 false)
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK1-NEXT:    [[TMP39:%.*]] = load float, float* [[TMP3]], align 4
// CHECK1-NEXT:    [[TMP40:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[CMP14:%.*]] = fcmp olt float [[TMP39]], [[TMP40]]
// CHECK1-NEXT:    br i1 [[CMP14]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1:       cond.true:
// CHECK1-NEXT:    [[TMP41:%.*]] = load float, float* [[TMP3]], align 4
// CHECK1-NEXT:    br label [[COND_END:%.*]]
// CHECK1:       cond.false:
// CHECK1-NEXT:    [[TMP42:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    br label [[COND_END]]
// CHECK1:       cond.end:
// CHECK1-NEXT:    [[COND:%.*]] = phi float [ [[TMP41]], [[COND_TRUE]] ], [ [[TMP42]], [[COND_FALSE]] ]
// CHECK1-NEXT:    store float [[COND]], float* [[TMP3]], align 4
// CHECK1-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1:       .omp.reduction.case2:
// CHECK1-NEXT:    [[TMP43:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[TMP44:%.*]] = bitcast float* [[TMP0]] to i32*
// CHECK1-NEXT:    [[ATOMIC_LOAD:%.*]] = load atomic i32, i32* [[TMP44]] monotonic, align 4
// CHECK1-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK1:       atomic_cont:
// CHECK1-NEXT:    [[TMP45:%.*]] = phi i32 [ [[ATOMIC_LOAD]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[TMP53:%.*]], [[ATOMIC_CONT]] ]
// CHECK1-NEXT:    [[TMP46:%.*]] = bitcast float* [[ATOMIC_TEMP]] to i32*
// CHECK1-NEXT:    [[TMP47:%.*]] = bitcast i32 [[TMP45]] to float
// CHECK1-NEXT:    store float [[TMP47]], float* [[TMP]], align 4
// CHECK1-NEXT:    [[TMP48:%.*]] = load float, float* [[TMP]], align 4
// CHECK1-NEXT:    [[TMP49:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[ADD15:%.*]] = fadd float [[TMP48]], [[TMP49]]
// CHECK1-NEXT:    store float [[ADD15]], float* [[ATOMIC_TEMP]], align 4
// CHECK1-NEXT:    [[TMP50:%.*]] = load i32, i32* [[TMP46]], align 4
// CHECK1-NEXT:    [[TMP51:%.*]] = bitcast float* [[TMP0]] to i32*
// CHECK1-NEXT:    [[TMP52:%.*]] = cmpxchg i32* [[TMP51]], i32 [[TMP45]], i32 [[TMP50]] monotonic monotonic, align 4
// CHECK1-NEXT:    [[TMP53]] = extractvalue { i32, i1 } [[TMP52]], 0
// CHECK1-NEXT:    [[TMP54:%.*]] = extractvalue { i32, i1 } [[TMP52]], 1
// CHECK1-NEXT:    br i1 [[TMP54]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK1:       atomic_exit:
// CHECK1-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    [[CALL16:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SIfEanERKS0_(%struct.S* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT:    [[TMP55:%.*]] = bitcast %struct.S* [[TMP1]] to i8*
// CHECK1-NEXT:    [[TMP56:%.*]] = bitcast %struct.S* [[CALL16]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP55]], i8* align 4 [[TMP56]], i64 4, i1 false)
// CHECK1-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    [[CALL18:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK1-NEXT:    [[TOBOOL19:%.*]] = fcmp une float [[CALL18]], 0.000000e+00
// CHECK1-NEXT:    br i1 [[TOBOOL19]], label [[LAND_RHS20:%.*]], label [[LAND_END23:%.*]]
// CHECK1:       land.rhs20:
// CHECK1-NEXT:    [[CALL21:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK1-NEXT:    [[TOBOOL22:%.*]] = fcmp une float [[CALL21]], 0.000000e+00
// CHECK1-NEXT:    br label [[LAND_END23]]
// CHECK1:       land.end23:
// CHECK1-NEXT:    [[TMP57:%.*]] = phi i1 [ false, [[ATOMIC_EXIT]] ], [ [[TOBOOL22]], [[LAND_RHS20]] ]
// CHECK1-NEXT:    [[CONV24:%.*]] = uitofp i1 [[TMP57]] to float
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP17]], float [[CONV24]])
// CHECK1-NEXT:    [[TMP58:%.*]] = bitcast %struct.S* [[TMP2]] to i8*
// CHECK1-NEXT:    [[TMP59:%.*]] = bitcast %struct.S* [[REF_TMP17]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP58]], i8* align 4 [[TMP59]], i64 4, i1 false)
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP17]]) #[[ATTR4]]
// CHECK1-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    [[TMP60:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[TMP61:%.*]] = bitcast float* [[TMP3]] to i32*
// CHECK1-NEXT:    [[ATOMIC_LOAD25:%.*]] = load atomic i32, i32* [[TMP61]] monotonic, align 4
// CHECK1-NEXT:    br label [[ATOMIC_CONT26:%.*]]
// CHECK1:       atomic_cont26:
// CHECK1-NEXT:    [[TMP62:%.*]] = phi i32 [ [[ATOMIC_LOAD25]], [[LAND_END23]] ], [ [[TMP72:%.*]], [[COND_END32:%.*]] ]
// CHECK1-NEXT:    [[TMP63:%.*]] = bitcast float* [[ATOMIC_TEMP27]] to i32*
// CHECK1-NEXT:    [[TMP64:%.*]] = bitcast i32 [[TMP62]] to float
// CHECK1-NEXT:    store float [[TMP64]], float* [[_TMP28]], align 4
// CHECK1-NEXT:    [[TMP65:%.*]] = load float, float* [[_TMP28]], align 4
// CHECK1-NEXT:    [[TMP66:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[CMP29:%.*]] = fcmp olt float [[TMP65]], [[TMP66]]
// CHECK1-NEXT:    br i1 [[CMP29]], label [[COND_TRUE30:%.*]], label [[COND_FALSE31:%.*]]
// CHECK1:       cond.true30:
// CHECK1-NEXT:    [[TMP67:%.*]] = load float, float* [[_TMP28]], align 4
// CHECK1-NEXT:    br label [[COND_END32]]
// CHECK1:       cond.false31:
// CHECK1-NEXT:    [[TMP68:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK1-NEXT:    br label [[COND_END32]]
// CHECK1:       cond.end32:
// CHECK1-NEXT:    [[COND33:%.*]] = phi float [ [[TMP67]], [[COND_TRUE30]] ], [ [[TMP68]], [[COND_FALSE31]] ]
// CHECK1-NEXT:    store float [[COND33]], float* [[ATOMIC_TEMP27]], align 4
// CHECK1-NEXT:    [[TMP69:%.*]] = load i32, i32* [[TMP63]], align 4
// CHECK1-NEXT:    [[TMP70:%.*]] = bitcast float* [[TMP3]] to i32*
// CHECK1-NEXT:    [[TMP71:%.*]] = cmpxchg i32* [[TMP70]], i32 [[TMP62]], i32 [[TMP69]] monotonic monotonic, align 4
// CHECK1-NEXT:    [[TMP72]] = extractvalue { i32, i1 } [[TMP71]], 0
// CHECK1-NEXT:    [[TMP73:%.*]] = extractvalue { i32, i1 } [[TMP71]], 1
// CHECK1-NEXT:    br i1 [[TMP73]], label [[ATOMIC_EXIT34:%.*]], label [[ATOMIC_CONT26]]
// CHECK1:       atomic_exit34:
// CHECK1-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1:       .omp.reduction.default:
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]]) #[[ATTR4]]
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR3]]) #[[ATTR4]]
// CHECK1-NEXT:    call void @__kmpc_barrier(%struct.ident_t* @[[GLOB4:[0-9]+]], i32 [[TMP7]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp.reduction.reduction_func
// CHECK1-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK1-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP2]] to [4 x i8*]*
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to [4 x i8*]*
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to float*
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP10:%.*]] = load i8*, i8** [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to float*
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[TMP12]], align 8
// CHECK1-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to %struct.S*
// CHECK1-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP16:%.*]] = load i8*, i8** [[TMP15]], align 8
// CHECK1-NEXT:    [[TMP17:%.*]] = bitcast i8* [[TMP16]] to %struct.S*
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = bitcast i8* [[TMP19]] to %struct.S*
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP22:%.*]] = load i8*, i8** [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = bitcast i8* [[TMP22]] to %struct.S*
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 3
// CHECK1-NEXT:    [[TMP25:%.*]] = load i8*, i8** [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = bitcast i8* [[TMP25]] to float*
// CHECK1-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 3
// CHECK1-NEXT:    [[TMP28:%.*]] = load i8*, i8** [[TMP27]], align 8
// CHECK1-NEXT:    [[TMP29:%.*]] = bitcast i8* [[TMP28]] to float*
// CHECK1-NEXT:    [[TMP30:%.*]] = load float, float* [[TMP11]], align 4
// CHECK1-NEXT:    [[TMP31:%.*]] = load float, float* [[TMP8]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = fadd float [[TMP30]], [[TMP31]]
// CHECK1-NEXT:    store float [[ADD]], float* [[TMP11]], align 4
// CHECK1-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SIfEanERKS0_(%struct.S* nonnull align 4 dereferenceable(4) [[TMP17]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP14]])
// CHECK1-NEXT:    [[TMP32:%.*]] = bitcast %struct.S* [[TMP17]] to i8*
// CHECK1-NEXT:    [[TMP33:%.*]] = bitcast %struct.S* [[CALL]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP32]], i8* align 4 [[TMP33]], i64 4, i1 false)
// CHECK1-NEXT:    [[CALL2:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP23]])
// CHECK1-NEXT:    [[TOBOOL:%.*]] = fcmp une float [[CALL2]], 0.000000e+00
// CHECK1-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK1:       land.rhs:
// CHECK1-NEXT:    [[CALL3:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP20]])
// CHECK1-NEXT:    [[TOBOOL4:%.*]] = fcmp une float [[CALL3]], 0.000000e+00
// CHECK1-NEXT:    br label [[LAND_END]]
// CHECK1:       land.end:
// CHECK1-NEXT:    [[TMP34:%.*]] = phi i1 [ false, [[ENTRY:%.*]] ], [ [[TOBOOL4]], [[LAND_RHS]] ]
// CHECK1-NEXT:    [[CONV:%.*]] = uitofp i1 [[TMP34]] to float
// CHECK1-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]], float [[CONV]])
// CHECK1-NEXT:    [[TMP35:%.*]] = bitcast %struct.S* [[TMP23]] to i8*
// CHECK1-NEXT:    [[TMP36:%.*]] = bitcast %struct.S* [[REF_TMP]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP35]], i8* align 4 [[TMP36]], i64 4, i1 false)
// CHECK1-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK1-NEXT:    [[TMP37:%.*]] = load float, float* [[TMP29]], align 4
// CHECK1-NEXT:    [[TMP38:%.*]] = load float, float* [[TMP26]], align 4
// CHECK1-NEXT:    [[CMP:%.*]] = fcmp olt float [[TMP37]], [[TMP38]]
// CHECK1-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1:       cond.true:
// CHECK1-NEXT:    [[TMP39:%.*]] = load float, float* [[TMP29]], align 4
// CHECK1-NEXT:    br label [[COND_END:%.*]]
// CHECK1:       cond.false:
// CHECK1-NEXT:    [[TMP40:%.*]] = load float, float* [[TMP26]], align 4
// CHECK1-NEXT:    br label [[COND_END]]
// CHECK1:       cond.end:
// CHECK1-NEXT:    [[COND:%.*]] = phi float [ [[TMP39]], [[COND_TRUE]] ], [ [[TMP40]], [[COND_FALSE]] ]
// CHECK1-NEXT:    store float [[COND]], float* [[TMP29]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfEanERKS0_
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR6:[0-9]+]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S* [[TMP0]], %struct.S** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret %struct.S* [[THIS1]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfEcvfEv
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) #[[ATTR6]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret float 0.000000e+00
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfED1Ev
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SIfED2Ev(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]]) #[[ATTR4]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_Z5tmainIiET_v
// CHECK1-SAME: () #[[ATTR6]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[T:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TEST:%.*]] = alloca [[STRUCT_S_0:%.*]], align 4
// CHECK1-NEXT:    [[T_VAR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[T_VAR1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[VEC:%.*]] = alloca [2 x i32], align 4
// CHECK1-NEXT:    [[S_ARR:%.*]] = alloca [2 x %struct.S.0], align 4
// CHECK1-NEXT:    [[VAR:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK1-NEXT:    [[VAR1:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[TEST]])
// CHECK1-NEXT:    store i32 0, i32* [[T_VAR]], align 4
// CHECK1-NEXT:    [[TMP0:%.*]] = bitcast [2 x i32]* [[VEC]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP0]], i8* align 4 bitcast ([2 x i32]* @__const._Z5tmainIiET_v.vec to i8*), i64 8, i1 false)
// CHECK1-NEXT:    [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], [2 x %struct.S.0]* [[S_ARR]], i64 0, i64 0
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 1)
// CHECK1-NEXT:    [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], %struct.S.0* [[ARRAYINIT_BEGIN]], i64 1
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 2)
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR]], i32 3)
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR1]])
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB3]], i32 6, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, %struct.S.0*, %struct.S.0*, i32*, [2 x i32]*, [2 x %struct.S.0]*)* @.omp_outlined..1 to void (i32*, i32*, ...)*), i32* [[T_VAR]], %struct.S.0* [[VAR]], %struct.S.0* [[VAR1]], i32* [[T_VAR1]], [2 x i32]* [[VEC]], [2 x %struct.S.0]* [[S_ARR]])
// CHECK1-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR1]]) #[[ATTR4]]
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR]]) #[[ATTR4]]
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], [2 x %struct.S.0]* [[S_ARR]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_S_0]], %struct.S.0* [[ARRAY_BEGIN]], i64 2
// CHECK1-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1:       arraydestroy.body:
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S.0* [ [[TMP1]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S_0]], %struct.S.0* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR4]]
// CHECK1-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S.0* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN]]
// CHECK1-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE1:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK1:       arraydestroy.done1:
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[TEST]]) #[[ATTR4]]
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    ret i32 [[TMP2]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfEC2Ev
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP0:%.*]] = load volatile double, double* @g, align 8
// CHECK1-NEXT:    [[CONV:%.*]] = fptrunc double [[TMP0]] to float
// CHECK1-NEXT:    store float [[CONV]], float* [[F]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfEC2Ef
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], float [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca float, align 4
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store float [[A]], float* [[A_ADDR]], align 4
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP0:%.*]] = load float, float* [[A_ADDR]], align 4
// CHECK1-NEXT:    [[CONV:%.*]] = fpext float [[TMP0]] to double
// CHECK1-NEXT:    [[TMP1:%.*]] = load volatile double, double* @g, align 8
// CHECK1-NEXT:    [[ADD:%.*]] = fadd double [[CONV]], [[TMP1]]
// CHECK1-NEXT:    [[CONV2:%.*]] = fptrunc double [[ADD]] to float
// CHECK1-NEXT:    store float [[CONV2]], float* [[F]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIfED2Ev
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiEC1Ev
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SIiEC2Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiEC1Ei
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
// CHECK1-NEXT:    call void @_ZN1SIiEC2Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS1]], i32 [[TMP0]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..1
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[T_VAR:%.*]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR:%.*]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR1:%.*]], i32* nonnull align 4 dereferenceable(4) [[T_VAR1:%.*]], [2 x i32]* nonnull align 4 dereferenceable(8) [[VEC:%.*]], [2 x %struct.S.0]* nonnull align 4 dereferenceable(8) [[S_ARR:%.*]]) #[[ATTR3]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[T_VAR_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[VAR_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    [[VAR1_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    [[T_VAR1_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[VEC_ADDR:%.*]] = alloca [2 x i32]*, align 8
// CHECK1-NEXT:    [[S_ARR_ADDR:%.*]] = alloca [2 x %struct.S.0]*, align 8
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_LB_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_UB_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_ST_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_IL_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_SECTIONS_IV_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[T_VAR2:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[VAR3:%.*]] = alloca [[STRUCT_S_0:%.*]], align 4
// CHECK1-NEXT:    [[VAR14:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK1-NEXT:    [[T_VAR15:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [4 x i8*], align 8
// CHECK1-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK1-NEXT:    [[REF_TMP13:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK1-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT:    store i32* [[T_VAR]], i32** [[T_VAR_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S.0* [[VAR]], %struct.S.0** [[VAR_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S.0* [[VAR1]], %struct.S.0** [[VAR1_ADDR]], align 8
// CHECK1-NEXT:    store i32* [[T_VAR1]], i32** [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT:    store [2 x i32]* [[VEC]], [2 x i32]** [[VEC_ADDR]], align 8
// CHECK1-NEXT:    store [2 x %struct.S.0]* [[S_ARR]], [2 x %struct.S.0]** [[S_ARR_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[T_VAR_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load %struct.S.0*, %struct.S.0** [[VAR_ADDR]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load %struct.S.0*, %struct.S.0** [[VAR1_ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32*, i32** [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load [2 x i32]*, [2 x i32]** [[VEC_ADDR]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = load [2 x %struct.S.0]*, [2 x %struct.S.0]** [[S_ARR_ADDR]], align 8
// CHECK1-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK1-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_ST_]], align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_IL_]], align 4
// CHECK1-NEXT:    store i32 0, i32* [[T_VAR2]], align 4
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK1-NEXT:    store i32 2147483647, i32* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[TMP6:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4
// CHECK1-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB1]], i32 [[TMP7]], i32 34, i32* [[DOTOMP_SECTIONS_IL_]], i32* [[DOTOMP_SECTIONS_LB_]], i32* [[DOTOMP_SECTIONS_UB_]], i32* [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
// CHECK1-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    [[TMP9:%.*]] = icmp slt i32 [[TMP8]], 1
// CHECK1-NEXT:    [[TMP10:%.*]] = select i1 [[TMP9]], i32 [[TMP8]], i32 1
// CHECK1-NEXT:    store i32 [[TMP10]], i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    [[TMP11:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK1-NEXT:    store i32 [[TMP11]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1:       omp.inner.for.cond:
// CHECK1-NEXT:    [[TMP12:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    [[TMP13:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT:    [[CMP:%.*]] = icmp sle i32 [[TMP12]], [[TMP13]]
// CHECK1-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1:       omp.inner.for.body:
// CHECK1-NEXT:    [[TMP14:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    switch i32 [[TMP14]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK1-NEXT:    i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK1-NEXT:    i32 1, label [[DOTOMP_SECTIONS_CASE6:%.*]]
// CHECK1-NEXT:    ]
// CHECK1:       .omp.sections.case:
// CHECK1-NEXT:    [[TMP15:%.*]] = load i32, i32* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[TMP4]], i64 0, i64 0
// CHECK1-NEXT:    store i32 [[TMP15]], i32* [[ARRAYIDX]], align 4
// CHECK1-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK1:       .omp.sections.case6:
// CHECK1-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds [2 x %struct.S.0], [2 x %struct.S.0]* [[TMP5]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP16:%.*]] = bitcast %struct.S.0* [[ARRAYIDX7]] to i8*
// CHECK1-NEXT:    [[TMP17:%.*]] = bitcast %struct.S.0* [[VAR3]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP16]], i8* align 4 [[TMP17]], i64 4, i1 false)
// CHECK1-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK1:       .omp.sections.exit:
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1:       omp.inner.for.inc:
// CHECK1-NEXT:    [[TMP18:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP18]], 1
// CHECK1-NEXT:    store i32 [[INC]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK1:       omp.inner.for.end:
// CHECK1-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[TMP7]])
// CHECK1-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP20:%.*]] = bitcast i32* [[T_VAR2]] to i8*
// CHECK1-NEXT:    store i8* [[TMP20]], i8** [[TMP19]], align 8
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP22:%.*]] = bitcast %struct.S.0* [[VAR3]] to i8*
// CHECK1-NEXT:    store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP24:%.*]] = bitcast %struct.S.0* [[VAR14]] to i8*
// CHECK1-NEXT:    store i8* [[TMP24]], i8** [[TMP23]], align 8
// CHECK1-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
// CHECK1-NEXT:    [[TMP26:%.*]] = bitcast i32* [[T_VAR15]] to i8*
// CHECK1-NEXT:    store i8* [[TMP26]], i8** [[TMP25]], align 8
// CHECK1-NEXT:    [[TMP27:%.*]] = bitcast [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT:    [[TMP28:%.*]] = call i32 @__kmpc_reduce_nowait(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], i32 4, i64 32, i8* [[TMP27]], void (i8*, i8*)* @.omp.reduction.reduction_func.2, [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT:    switch i32 [[TMP28]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK1-NEXT:    i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK1-NEXT:    i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK1-NEXT:    ]
// CHECK1:       .omp.reduction.case1:
// CHECK1-NEXT:    [[TMP29:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[TMP30:%.*]] = load i32, i32* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP29]], [[TMP30]]
// CHECK1-NEXT:    store i32 [[ADD]], i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S.0* @_ZN1SIiEanERKS0_(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT:    [[TMP31:%.*]] = bitcast %struct.S.0* [[TMP1]] to i8*
// CHECK1-NEXT:    [[TMP32:%.*]] = bitcast %struct.S.0* [[CALL]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP31]], i8* align 4 [[TMP32]], i64 4, i1 false)
// CHECK1-NEXT:    [[CALL8:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK1-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[CALL8]], 0
// CHECK1-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK1:       land.rhs:
// CHECK1-NEXT:    [[CALL9:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK1-NEXT:    [[TOBOOL10:%.*]] = icmp ne i32 [[CALL9]], 0
// CHECK1-NEXT:    br label [[LAND_END]]
// CHECK1:       land.end:
// CHECK1-NEXT:    [[TMP33:%.*]] = phi i1 [ false, [[DOTOMP_REDUCTION_CASE1]] ], [ [[TOBOOL10]], [[LAND_RHS]] ]
// CHECK1-NEXT:    [[CONV:%.*]] = zext i1 [[TMP33]] to i32
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]], i32 [[CONV]])
// CHECK1-NEXT:    [[TMP34:%.*]] = bitcast %struct.S.0* [[TMP2]] to i8*
// CHECK1-NEXT:    [[TMP35:%.*]] = bitcast %struct.S.0* [[REF_TMP]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP34]], i8* align 4 [[TMP35]], i64 4, i1 false)
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK1-NEXT:    [[TMP36:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK1-NEXT:    [[TMP37:%.*]] = load i32, i32* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[CMP11:%.*]] = icmp slt i32 [[TMP36]], [[TMP37]]
// CHECK1-NEXT:    br i1 [[CMP11]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1:       cond.true:
// CHECK1-NEXT:    [[TMP38:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK1-NEXT:    br label [[COND_END:%.*]]
// CHECK1:       cond.false:
// CHECK1-NEXT:    [[TMP39:%.*]] = load i32, i32* [[T_VAR15]], align 4
// CHECK1-NEXT:    br label [[COND_END]]
// CHECK1:       cond.end:
// CHECK1-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP38]], [[COND_TRUE]] ], [ [[TMP39]], [[COND_FALSE]] ]
// CHECK1-NEXT:    store i32 [[COND]], i32* [[TMP3]], align 4
// CHECK1-NEXT:    call void @__kmpc_end_reduce_nowait(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1:       .omp.reduction.case2:
// CHECK1-NEXT:    [[TMP40:%.*]] = load i32, i32* [[T_VAR2]], align 4
// CHECK1-NEXT:    [[TMP41:%.*]] = atomicrmw add i32* [[TMP0]], i32 [[TMP40]] monotonic, align 4
// CHECK1-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    [[CALL12:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S.0* @_ZN1SIiEanERKS0_(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT:    [[TMP42:%.*]] = bitcast %struct.S.0* [[TMP1]] to i8*
// CHECK1-NEXT:    [[TMP43:%.*]] = bitcast %struct.S.0* [[CALL12]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP42]], i8* align 4 [[TMP43]], i64 4, i1 false)
// CHECK1-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    [[CALL14:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK1-NEXT:    [[TOBOOL15:%.*]] = icmp ne i32 [[CALL14]], 0
// CHECK1-NEXT:    br i1 [[TOBOOL15]], label [[LAND_RHS16:%.*]], label [[LAND_END19:%.*]]
// CHECK1:       land.rhs16:
// CHECK1-NEXT:    [[CALL17:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK1-NEXT:    [[TOBOOL18:%.*]] = icmp ne i32 [[CALL17]], 0
// CHECK1-NEXT:    br label [[LAND_END19]]
// CHECK1:       land.end19:
// CHECK1-NEXT:    [[TMP44:%.*]] = phi i1 [ false, [[DOTOMP_REDUCTION_CASE2]] ], [ [[TOBOOL18]], [[LAND_RHS16]] ]
// CHECK1-NEXT:    [[CONV20:%.*]] = zext i1 [[TMP44]] to i32
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP13]], i32 [[CONV20]])
// CHECK1-NEXT:    [[TMP45:%.*]] = bitcast %struct.S.0* [[TMP2]] to i8*
// CHECK1-NEXT:    [[TMP46:%.*]] = bitcast %struct.S.0* [[REF_TMP13]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP45]], i8* align 4 [[TMP46]], i64 4, i1 false)
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP13]]) #[[ATTR4]]
// CHECK1-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT:    [[TMP47:%.*]] = load i32, i32* [[T_VAR15]], align 4
// CHECK1-NEXT:    [[TMP48:%.*]] = atomicrmw min i32* [[TMP3]], i32 [[TMP47]] monotonic, align 4
// CHECK1-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1:       .omp.reduction.default:
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]]) #[[ATTR4]]
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]]) #[[ATTR4]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp.reduction.reduction_func.2
// CHECK1-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S_0:%.*]], align 4
// CHECK1-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP2]] to [4 x i8*]*
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to [4 x i8*]*
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to i32*
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP10:%.*]] = load i8*, i8** [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to i32*
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[TMP12]], align 8
// CHECK1-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to %struct.S.0*
// CHECK1-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP16:%.*]] = load i8*, i8** [[TMP15]], align 8
// CHECK1-NEXT:    [[TMP17:%.*]] = bitcast i8* [[TMP16]] to %struct.S.0*
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = bitcast i8* [[TMP19]] to %struct.S.0*
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP22:%.*]] = load i8*, i8** [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = bitcast i8* [[TMP22]] to %struct.S.0*
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 3
// CHECK1-NEXT:    [[TMP25:%.*]] = load i8*, i8** [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = bitcast i8* [[TMP25]] to i32*
// CHECK1-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 3
// CHECK1-NEXT:    [[TMP28:%.*]] = load i8*, i8** [[TMP27]], align 8
// CHECK1-NEXT:    [[TMP29:%.*]] = bitcast i8* [[TMP28]] to i32*
// CHECK1-NEXT:    [[TMP30:%.*]] = load i32, i32* [[TMP11]], align 4
// CHECK1-NEXT:    [[TMP31:%.*]] = load i32, i32* [[TMP8]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP30]], [[TMP31]]
// CHECK1-NEXT:    store i32 [[ADD]], i32* [[TMP11]], align 4
// CHECK1-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S.0* @_ZN1SIiEanERKS0_(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP17]], %struct.S.0* nonnull align 4 dereferenceable(4) [[TMP14]])
// CHECK1-NEXT:    [[TMP32:%.*]] = bitcast %struct.S.0* [[TMP17]] to i8*
// CHECK1-NEXT:    [[TMP33:%.*]] = bitcast %struct.S.0* [[CALL]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP32]], i8* align 4 [[TMP33]], i64 4, i1 false)
// CHECK1-NEXT:    [[CALL2:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP23]])
// CHECK1-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[CALL2]], 0
// CHECK1-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK1:       land.rhs:
// CHECK1-NEXT:    [[CALL3:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP20]])
// CHECK1-NEXT:    [[TOBOOL4:%.*]] = icmp ne i32 [[CALL3]], 0
// CHECK1-NEXT:    br label [[LAND_END]]
// CHECK1:       land.end:
// CHECK1-NEXT:    [[TMP34:%.*]] = phi i1 [ false, [[ENTRY:%.*]] ], [ [[TOBOOL4]], [[LAND_RHS]] ]
// CHECK1-NEXT:    [[CONV:%.*]] = zext i1 [[TMP34]] to i32
// CHECK1-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]], i32 [[CONV]])
// CHECK1-NEXT:    [[TMP35:%.*]] = bitcast %struct.S.0* [[TMP23]] to i8*
// CHECK1-NEXT:    [[TMP36:%.*]] = bitcast %struct.S.0* [[REF_TMP]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP35]], i8* align 4 [[TMP36]], i64 4, i1 false)
// CHECK1-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK1-NEXT:    [[TMP37:%.*]] = load i32, i32* [[TMP29]], align 4
// CHECK1-NEXT:    [[TMP38:%.*]] = load i32, i32* [[TMP26]], align 4
// CHECK1-NEXT:    [[CMP:%.*]] = icmp slt i32 [[TMP37]], [[TMP38]]
// CHECK1-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1:       cond.true:
// CHECK1-NEXT:    [[TMP39:%.*]] = load i32, i32* [[TMP29]], align 4
// CHECK1-NEXT:    br label [[COND_END:%.*]]
// CHECK1:       cond.false:
// CHECK1-NEXT:    [[TMP40:%.*]] = load i32, i32* [[TMP26]], align 4
// CHECK1-NEXT:    br label [[COND_END]]
// CHECK1:       cond.end:
// CHECK1-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP39]], [[COND_TRUE]] ], [ [[TMP40]], [[COND_FALSE]] ]
// CHECK1-NEXT:    store i32 [[COND]], i32* [[TMP29]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiEanERKS0_
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]], %struct.S.0* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR6]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S.0* [[TMP0]], %struct.S.0** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret %struct.S.0* [[THIS1]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiEcviEv
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) #[[ATTR6]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiED1Ev
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SIiED2Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS1]]) #[[ATTR4]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiEC2Ev
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S_0:%.*]], %struct.S.0* [[THIS1]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP0:%.*]] = load volatile double, double* @g, align 8
// CHECK1-NEXT:    [[CONV:%.*]] = fptosi double [[TMP0]] to i32
// CHECK1-NEXT:    store i32 [[CONV]], i32* [[F]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiEC2Ei
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S_0:%.*]], %struct.S.0* [[THIS1]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
// CHECK1-NEXT:    [[CONV:%.*]] = sitofp i32 [[TMP0]] to double
// CHECK1-NEXT:    [[TMP1:%.*]] = load volatile double, double* @g, align 8
// CHECK1-NEXT:    [[ADD:%.*]] = fadd double [[CONV]], [[TMP1]]
// CHECK1-NEXT:    [[CONV2:%.*]] = fptosi double [[ADD]] to i32
// CHECK1-NEXT:    store i32 [[CONV2]], i32* [[F]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SIiED2Ev
// CHECK1-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK1-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@main
// CHECK2-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TEST:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK2-NEXT:    [[T_VAR:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[T_VAR1:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[VEC:%.*]] = alloca [2 x i32], align 4
// CHECK2-NEXT:    [[S_ARR:%.*]] = alloca [2 x %struct.S], align 4
// CHECK2-NEXT:    [[VAR:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK2-NEXT:    [[VAR1:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK2-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[TEST]])
// CHECK2-NEXT:    store float 0.000000e+00, float* [[T_VAR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = bitcast [2 x i32]* [[VEC]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP0]], i8* align 4 bitcast ([2 x i32]* @__const.main.vec to i8*), i64 8, i1 false)
// CHECK2-NEXT:    [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[S_ARR]], i64 0, i64 0
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], float 1.000000e+00)
// CHECK2-NEXT:    [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYINIT_BEGIN]], i64 1
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], float 2.000000e+00)
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[VAR]], float 3.000000e+00)
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR1]])
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB3:[0-9]+]], i32 6, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float*, %struct.S*, %struct.S*, float*, [2 x i32]*, [2 x %struct.S]*)* @.omp_outlined. to void (i32*, i32*, ...)*), float* [[T_VAR]], %struct.S* [[VAR]], %struct.S* [[VAR1]], float* [[T_VAR1]], [2 x i32]* [[VEC]], [2 x %struct.S]* [[S_ARR]])
// CHECK2-NEXT:    [[CALL:%.*]] = call i32 @_Z5tmainIiET_v()
// CHECK2-NEXT:    store i32 [[CALL]], i32* [[RETVAL]], align 4
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR1]]) #[[ATTR4:[0-9]+]]
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR]]) #[[ATTR4]]
// CHECK2-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[S_ARR]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAY_BEGIN]], i64 2
// CHECK2-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK2:       arraydestroy.body:
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP1]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR4]]
// CHECK2-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN]]
// CHECK2-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE1:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK2:       arraydestroy.done1:
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[TEST]]) #[[ATTR4]]
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    ret i32 [[TMP2]]
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfEC1Ev
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1:[0-9]+]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    call void @_ZN1SIfEC2Ev(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfEC1Ef
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], float [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca float, align 4
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store float [[A]], float* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load float, float* [[A_ADDR]], align 4
// CHECK2-NEXT:    call void @_ZN1SIfEC2Ef(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]], float [[TMP0]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], float* nonnull align 4 dereferenceable(4) [[T_VAR:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR1:%.*]], float* nonnull align 4 dereferenceable(4) [[T_VAR1:%.*]], [2 x i32]* nonnull align 4 dereferenceable(8) [[VEC:%.*]], [2 x %struct.S]* nonnull align 4 dereferenceable(8) [[S_ARR:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[T_VAR_ADDR:%.*]] = alloca float*, align 8
// CHECK2-NEXT:    [[VAR_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[VAR1_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[T_VAR1_ADDR:%.*]] = alloca float*, align 8
// CHECK2-NEXT:    [[VEC_ADDR:%.*]] = alloca [2 x i32]*, align 8
// CHECK2-NEXT:    [[S_ARR_ADDR:%.*]] = alloca [2 x %struct.S]*, align 8
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_LB_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_UB_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_ST_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_IL_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_IV_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[T_VAR2:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[VAR3:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK2-NEXT:    [[VAR14:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK2-NEXT:    [[T_VAR15:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [4 x i8*], align 8
// CHECK2-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK2-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[TMP:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[REF_TMP17:%.*]] = alloca [[STRUCT_S]], align 4
// CHECK2-NEXT:    [[ATOMIC_TEMP27:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[_TMP28:%.*]] = alloca float, align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
// CHECK2-NEXT:    store float* [[T_VAR]], float** [[T_VAR_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S* [[VAR]], %struct.S** [[VAR_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S* [[VAR1]], %struct.S** [[VAR1_ADDR]], align 8
// CHECK2-NEXT:    store float* [[T_VAR1]], float** [[T_VAR1_ADDR]], align 8
// CHECK2-NEXT:    store [2 x i32]* [[VEC]], [2 x i32]** [[VEC_ADDR]], align 8
// CHECK2-NEXT:    store [2 x %struct.S]* [[S_ARR]], [2 x %struct.S]** [[S_ARR_ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load float*, float** [[T_VAR_ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load %struct.S*, %struct.S** [[VAR_ADDR]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load %struct.S*, %struct.S** [[VAR1_ADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load float*, float** [[T_VAR1_ADDR]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = load [2 x i32]*, [2 x i32]** [[VEC_ADDR]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = load [2 x %struct.S]*, [2 x %struct.S]** [[S_ARR_ADDR]], align 8
// CHECK2-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_ST_]], align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_IL_]], align 4
// CHECK2-NEXT:    store float 0.000000e+00, float* [[T_VAR2]], align 4
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK2-NEXT:    store float 0x47EFFFFFE0000000, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4
// CHECK2-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB1:[0-9]+]], i32 [[TMP7]], i32 34, i32* [[DOTOMP_SECTIONS_IL_]], i32* [[DOTOMP_SECTIONS_LB_]], i32* [[DOTOMP_SECTIONS_UB_]], i32* [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
// CHECK2-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    [[TMP9:%.*]] = icmp slt i32 [[TMP8]], 0
// CHECK2-NEXT:    [[TMP10:%.*]] = select i1 [[TMP9]], i32 [[TMP8]], i32 0
// CHECK2-NEXT:    store i32 [[TMP10]], i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    [[TMP11:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK2-NEXT:    store i32 [[TMP11]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK2:       omp.inner.for.cond:
// CHECK2-NEXT:    [[TMP12:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    [[TMP13:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    [[CMP:%.*]] = icmp sle i32 [[TMP12]], [[TMP13]]
// CHECK2-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK2:       omp.inner.for.body:
// CHECK2-NEXT:    [[TMP14:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    switch i32 [[TMP14]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK2-NEXT:    i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK2-NEXT:    ]
// CHECK2:       .omp.sections.case:
// CHECK2-NEXT:    [[TMP15:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = fptosi float [[TMP15]] to i32
// CHECK2-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[TMP4]], i64 0, i64 0
// CHECK2-NEXT:    store i32 [[CONV]], i32* [[ARRAYIDX]], align 4
// CHECK2-NEXT:    [[ARRAYIDX6:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[TMP5]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP16:%.*]] = bitcast %struct.S* [[ARRAYIDX6]] to i8*
// CHECK2-NEXT:    [[TMP17:%.*]] = bitcast %struct.S* [[VAR3]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP16]], i8* align 4 [[TMP17]], i64 4, i1 false)
// CHECK2-NEXT:    [[TMP18:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[CONV7:%.*]] = fptosi float [[TMP18]] to i32
// CHECK2-NEXT:    [[ARRAYIDX8:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[TMP4]], i64 0, i64 1
// CHECK2-NEXT:    store i32 [[CONV7]], i32* [[ARRAYIDX8]], align 4
// CHECK2-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds [2 x %struct.S], [2 x %struct.S]* [[TMP5]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP19:%.*]] = bitcast %struct.S* [[ARRAYIDX9]] to i8*
// CHECK2-NEXT:    [[TMP20:%.*]] = bitcast %struct.S* [[VAR14]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP19]], i8* align 4 [[TMP20]], i64 4, i1 false)
// CHECK2-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK2:       .omp.sections.exit:
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK2:       omp.inner.for.inc:
// CHECK2-NEXT:    [[TMP21:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP21]], 1
// CHECK2-NEXT:    store i32 [[INC]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK2:       omp.inner.for.end:
// CHECK2-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[TMP7]])
// CHECK2-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP23:%.*]] = bitcast float* [[T_VAR2]] to i8*
// CHECK2-NEXT:    store i8* [[TMP23]], i8** [[TMP22]], align 8
// CHECK2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP25:%.*]] = bitcast %struct.S* [[VAR3]] to i8*
// CHECK2-NEXT:    store i8* [[TMP25]], i8** [[TMP24]], align 8
// CHECK2-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP27:%.*]] = bitcast %struct.S* [[VAR14]] to i8*
// CHECK2-NEXT:    store i8* [[TMP27]], i8** [[TMP26]], align 8
// CHECK2-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
// CHECK2-NEXT:    [[TMP29:%.*]] = bitcast float* [[T_VAR15]] to i8*
// CHECK2-NEXT:    store i8* [[TMP29]], i8** [[TMP28]], align 8
// CHECK2-NEXT:    [[TMP30:%.*]] = bitcast [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT:    [[TMP31:%.*]] = call i32 @__kmpc_reduce(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP7]], i32 4, i64 32, i8* [[TMP30]], void (i8*, i8*)* @.omp.reduction.reduction_func, [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK2-NEXT:    switch i32 [[TMP31]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK2-NEXT:    i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK2-NEXT:    i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK2-NEXT:    ]
// CHECK2:       .omp.reduction.case1:
// CHECK2-NEXT:    [[TMP32:%.*]] = load float, float* [[TMP0]], align 4
// CHECK2-NEXT:    [[TMP33:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[ADD:%.*]] = fadd float [[TMP32]], [[TMP33]]
// CHECK2-NEXT:    store float [[ADD]], float* [[TMP0]], align 4
// CHECK2-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SIfEanERKS0_(%struct.S* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK2-NEXT:    [[TMP34:%.*]] = bitcast %struct.S* [[TMP1]] to i8*
// CHECK2-NEXT:    [[TMP35:%.*]] = bitcast %struct.S* [[CALL]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP34]], i8* align 4 [[TMP35]], i64 4, i1 false)
// CHECK2-NEXT:    [[CALL10:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK2-NEXT:    [[TOBOOL:%.*]] = fcmp une float [[CALL10]], 0.000000e+00
// CHECK2-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK2:       land.rhs:
// CHECK2-NEXT:    [[CALL11:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK2-NEXT:    [[TOBOOL12:%.*]] = fcmp une float [[CALL11]], 0.000000e+00
// CHECK2-NEXT:    br label [[LAND_END]]
// CHECK2:       land.end:
// CHECK2-NEXT:    [[TMP36:%.*]] = phi i1 [ false, [[DOTOMP_REDUCTION_CASE1]] ], [ [[TOBOOL12]], [[LAND_RHS]] ]
// CHECK2-NEXT:    [[CONV13:%.*]] = uitofp i1 [[TMP36]] to float
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]], float [[CONV13]])
// CHECK2-NEXT:    [[TMP37:%.*]] = bitcast %struct.S* [[TMP2]] to i8*
// CHECK2-NEXT:    [[TMP38:%.*]] = bitcast %struct.S* [[REF_TMP]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP37]], i8* align 4 [[TMP38]], i64 4, i1 false)
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK2-NEXT:    [[TMP39:%.*]] = load float, float* [[TMP3]], align 4
// CHECK2-NEXT:    [[TMP40:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[CMP14:%.*]] = fcmp olt float [[TMP39]], [[TMP40]]
// CHECK2-NEXT:    br i1 [[CMP14]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK2:       cond.true:
// CHECK2-NEXT:    [[TMP41:%.*]] = load float, float* [[TMP3]], align 4
// CHECK2-NEXT:    br label [[COND_END:%.*]]
// CHECK2:       cond.false:
// CHECK2-NEXT:    [[TMP42:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    br label [[COND_END]]
// CHECK2:       cond.end:
// CHECK2-NEXT:    [[COND:%.*]] = phi float [ [[TMP41]], [[COND_TRUE]] ], [ [[TMP42]], [[COND_FALSE]] ]
// CHECK2-NEXT:    store float [[COND]], float* [[TMP3]], align 4
// CHECK2-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK2-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK2:       .omp.reduction.case2:
// CHECK2-NEXT:    [[TMP43:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[TMP44:%.*]] = bitcast float* [[TMP0]] to i32*
// CHECK2-NEXT:    [[ATOMIC_LOAD:%.*]] = load atomic i32, i32* [[TMP44]] monotonic, align 4
// CHECK2-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK2:       atomic_cont:
// CHECK2-NEXT:    [[TMP45:%.*]] = phi i32 [ [[ATOMIC_LOAD]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[TMP53:%.*]], [[ATOMIC_CONT]] ]
// CHECK2-NEXT:    [[TMP46:%.*]] = bitcast float* [[ATOMIC_TEMP]] to i32*
// CHECK2-NEXT:    [[TMP47:%.*]] = bitcast i32 [[TMP45]] to float
// CHECK2-NEXT:    store float [[TMP47]], float* [[TMP]], align 4
// CHECK2-NEXT:    [[TMP48:%.*]] = load float, float* [[TMP]], align 4
// CHECK2-NEXT:    [[TMP49:%.*]] = load float, float* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[ADD15:%.*]] = fadd float [[TMP48]], [[TMP49]]
// CHECK2-NEXT:    store float [[ADD15]], float* [[ATOMIC_TEMP]], align 4
// CHECK2-NEXT:    [[TMP50:%.*]] = load i32, i32* [[TMP46]], align 4
// CHECK2-NEXT:    [[TMP51:%.*]] = bitcast float* [[TMP0]] to i32*
// CHECK2-NEXT:    [[TMP52:%.*]] = cmpxchg i32* [[TMP51]], i32 [[TMP45]], i32 [[TMP50]] monotonic monotonic, align 4
// CHECK2-NEXT:    [[TMP53]] = extractvalue { i32, i1 } [[TMP52]], 0
// CHECK2-NEXT:    [[TMP54:%.*]] = extractvalue { i32, i1 } [[TMP52]], 1
// CHECK2-NEXT:    br i1 [[TMP54]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK2:       atomic_exit:
// CHECK2-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    [[CALL16:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SIfEanERKS0_(%struct.S* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK2-NEXT:    [[TMP55:%.*]] = bitcast %struct.S* [[TMP1]] to i8*
// CHECK2-NEXT:    [[TMP56:%.*]] = bitcast %struct.S* [[CALL16]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP55]], i8* align 4 [[TMP56]], i64 4, i1 false)
// CHECK2-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    [[CALL18:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK2-NEXT:    [[TOBOOL19:%.*]] = fcmp une float [[CALL18]], 0.000000e+00
// CHECK2-NEXT:    br i1 [[TOBOOL19]], label [[LAND_RHS20:%.*]], label [[LAND_END23:%.*]]
// CHECK2:       land.rhs20:
// CHECK2-NEXT:    [[CALL21:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK2-NEXT:    [[TOBOOL22:%.*]] = fcmp une float [[CALL21]], 0.000000e+00
// CHECK2-NEXT:    br label [[LAND_END23]]
// CHECK2:       land.end23:
// CHECK2-NEXT:    [[TMP57:%.*]] = phi i1 [ false, [[ATOMIC_EXIT]] ], [ [[TOBOOL22]], [[LAND_RHS20]] ]
// CHECK2-NEXT:    [[CONV24:%.*]] = uitofp i1 [[TMP57]] to float
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP17]], float [[CONV24]])
// CHECK2-NEXT:    [[TMP58:%.*]] = bitcast %struct.S* [[TMP2]] to i8*
// CHECK2-NEXT:    [[TMP59:%.*]] = bitcast %struct.S* [[REF_TMP17]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP58]], i8* align 4 [[TMP59]], i64 4, i1 false)
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP17]]) #[[ATTR4]]
// CHECK2-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    [[TMP60:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[TMP61:%.*]] = bitcast float* [[TMP3]] to i32*
// CHECK2-NEXT:    [[ATOMIC_LOAD25:%.*]] = load atomic i32, i32* [[TMP61]] monotonic, align 4
// CHECK2-NEXT:    br label [[ATOMIC_CONT26:%.*]]
// CHECK2:       atomic_cont26:
// CHECK2-NEXT:    [[TMP62:%.*]] = phi i32 [ [[ATOMIC_LOAD25]], [[LAND_END23]] ], [ [[TMP72:%.*]], [[COND_END32:%.*]] ]
// CHECK2-NEXT:    [[TMP63:%.*]] = bitcast float* [[ATOMIC_TEMP27]] to i32*
// CHECK2-NEXT:    [[TMP64:%.*]] = bitcast i32 [[TMP62]] to float
// CHECK2-NEXT:    store float [[TMP64]], float* [[_TMP28]], align 4
// CHECK2-NEXT:    [[TMP65:%.*]] = load float, float* [[_TMP28]], align 4
// CHECK2-NEXT:    [[TMP66:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[CMP29:%.*]] = fcmp olt float [[TMP65]], [[TMP66]]
// CHECK2-NEXT:    br i1 [[CMP29]], label [[COND_TRUE30:%.*]], label [[COND_FALSE31:%.*]]
// CHECK2:       cond.true30:
// CHECK2-NEXT:    [[TMP67:%.*]] = load float, float* [[_TMP28]], align 4
// CHECK2-NEXT:    br label [[COND_END32]]
// CHECK2:       cond.false31:
// CHECK2-NEXT:    [[TMP68:%.*]] = load float, float* [[T_VAR15]], align 4
// CHECK2-NEXT:    br label [[COND_END32]]
// CHECK2:       cond.end32:
// CHECK2-NEXT:    [[COND33:%.*]] = phi float [ [[TMP67]], [[COND_TRUE30]] ], [ [[TMP68]], [[COND_FALSE31]] ]
// CHECK2-NEXT:    store float [[COND33]], float* [[ATOMIC_TEMP27]], align 4
// CHECK2-NEXT:    [[TMP69:%.*]] = load i32, i32* [[TMP63]], align 4
// CHECK2-NEXT:    [[TMP70:%.*]] = bitcast float* [[TMP3]] to i32*
// CHECK2-NEXT:    [[TMP71:%.*]] = cmpxchg i32* [[TMP70]], i32 [[TMP62]], i32 [[TMP69]] monotonic monotonic, align 4
// CHECK2-NEXT:    [[TMP72]] = extractvalue { i32, i1 } [[TMP71]], 0
// CHECK2-NEXT:    [[TMP73:%.*]] = extractvalue { i32, i1 } [[TMP71]], 1
// CHECK2-NEXT:    br i1 [[TMP73]], label [[ATOMIC_EXIT34:%.*]], label [[ATOMIC_CONT26]]
// CHECK2:       atomic_exit34:
// CHECK2-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK2-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK2:       .omp.reduction.default:
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR14]]) #[[ATTR4]]
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[VAR3]]) #[[ATTR4]]
// CHECK2-NEXT:    call void @__kmpc_barrier(%struct.ident_t* @[[GLOB4:[0-9]+]], i32 [[TMP7]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp.reduction.reduction_func
// CHECK2-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S:%.*]], align 4
// CHECK2-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP2]] to [4 x i8*]*
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to [4 x i8*]*
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK2-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to float*
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP10:%.*]] = load i8*, i8** [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to float*
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[TMP12]], align 8
// CHECK2-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to %struct.S*
// CHECK2-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP16:%.*]] = load i8*, i8** [[TMP15]], align 8
// CHECK2-NEXT:    [[TMP17:%.*]] = bitcast i8* [[TMP16]] to %struct.S*
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    [[TMP20:%.*]] = bitcast i8* [[TMP19]] to %struct.S*
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP22:%.*]] = load i8*, i8** [[TMP21]], align 8
// CHECK2-NEXT:    [[TMP23:%.*]] = bitcast i8* [[TMP22]] to %struct.S*
// CHECK2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 3
// CHECK2-NEXT:    [[TMP25:%.*]] = load i8*, i8** [[TMP24]], align 8
// CHECK2-NEXT:    [[TMP26:%.*]] = bitcast i8* [[TMP25]] to float*
// CHECK2-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 3
// CHECK2-NEXT:    [[TMP28:%.*]] = load i8*, i8** [[TMP27]], align 8
// CHECK2-NEXT:    [[TMP29:%.*]] = bitcast i8* [[TMP28]] to float*
// CHECK2-NEXT:    [[TMP30:%.*]] = load float, float* [[TMP11]], align 4
// CHECK2-NEXT:    [[TMP31:%.*]] = load float, float* [[TMP8]], align 4
// CHECK2-NEXT:    [[ADD:%.*]] = fadd float [[TMP30]], [[TMP31]]
// CHECK2-NEXT:    store float [[ADD]], float* [[TMP11]], align 4
// CHECK2-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SIfEanERKS0_(%struct.S* nonnull align 4 dereferenceable(4) [[TMP17]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP14]])
// CHECK2-NEXT:    [[TMP32:%.*]] = bitcast %struct.S* [[TMP17]] to i8*
// CHECK2-NEXT:    [[TMP33:%.*]] = bitcast %struct.S* [[CALL]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP32]], i8* align 4 [[TMP33]], i64 4, i1 false)
// CHECK2-NEXT:    [[CALL2:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP23]])
// CHECK2-NEXT:    [[TOBOOL:%.*]] = fcmp une float [[CALL2]], 0.000000e+00
// CHECK2-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK2:       land.rhs:
// CHECK2-NEXT:    [[CALL3:%.*]] = call float @_ZN1SIfEcvfEv(%struct.S* nonnull align 4 dereferenceable(4) [[TMP20]])
// CHECK2-NEXT:    [[TOBOOL4:%.*]] = fcmp une float [[CALL3]], 0.000000e+00
// CHECK2-NEXT:    br label [[LAND_END]]
// CHECK2:       land.end:
// CHECK2-NEXT:    [[TMP34:%.*]] = phi i1 [ false, [[ENTRY:%.*]] ], [ [[TOBOOL4]], [[LAND_RHS]] ]
// CHECK2-NEXT:    [[CONV:%.*]] = uitofp i1 [[TMP34]] to float
// CHECK2-NEXT:    call void @_ZN1SIfEC1Ef(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]], float [[CONV]])
// CHECK2-NEXT:    [[TMP35:%.*]] = bitcast %struct.S* [[TMP23]] to i8*
// CHECK2-NEXT:    [[TMP36:%.*]] = bitcast %struct.S* [[REF_TMP]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP35]], i8* align 4 [[TMP36]], i64 4, i1 false)
// CHECK2-NEXT:    call void @_ZN1SIfED1Ev(%struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK2-NEXT:    [[TMP37:%.*]] = load float, float* [[TMP29]], align 4
// CHECK2-NEXT:    [[TMP38:%.*]] = load float, float* [[TMP26]], align 4
// CHECK2-NEXT:    [[CMP:%.*]] = fcmp olt float [[TMP37]], [[TMP38]]
// CHECK2-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK2:       cond.true:
// CHECK2-NEXT:    [[TMP39:%.*]] = load float, float* [[TMP29]], align 4
// CHECK2-NEXT:    br label [[COND_END:%.*]]
// CHECK2:       cond.false:
// CHECK2-NEXT:    [[TMP40:%.*]] = load float, float* [[TMP26]], align 4
// CHECK2-NEXT:    br label [[COND_END]]
// CHECK2:       cond.end:
// CHECK2-NEXT:    [[COND:%.*]] = phi float [ [[TMP39]], [[COND_TRUE]] ], [ [[TMP40]], [[COND_FALSE]] ]
// CHECK2-NEXT:    store float [[COND]], float* [[TMP29]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfEanERKS0_
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR6:[0-9]+]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S* [[TMP0]], %struct.S** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    ret %struct.S* [[THIS1]]
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfEcvfEv
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) #[[ATTR6]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    ret float 0.000000e+00
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfED1Ev
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    call void @_ZN1SIfED2Ev(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]]) #[[ATTR4]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_Z5tmainIiET_v
// CHECK2-SAME: () #[[ATTR6]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[T:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TEST:%.*]] = alloca [[STRUCT_S_0:%.*]], align 4
// CHECK2-NEXT:    [[T_VAR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[T_VAR1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[VEC:%.*]] = alloca [2 x i32], align 4
// CHECK2-NEXT:    [[S_ARR:%.*]] = alloca [2 x %struct.S.0], align 4
// CHECK2-NEXT:    [[VAR:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK2-NEXT:    [[VAR1:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[TEST]])
// CHECK2-NEXT:    store i32 0, i32* [[T_VAR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = bitcast [2 x i32]* [[VEC]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP0]], i8* align 4 bitcast ([2 x i32]* @__const._Z5tmainIiET_v.vec to i8*), i64 8, i1 false)
// CHECK2-NEXT:    [[ARRAYINIT_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], [2 x %struct.S.0]* [[S_ARR]], i64 0, i64 0
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[ARRAYINIT_BEGIN]], i32 1)
// CHECK2-NEXT:    [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [[STRUCT_S_0]], %struct.S.0* [[ARRAYINIT_BEGIN]], i64 1
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[ARRAYINIT_ELEMENT]], i32 2)
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR]], i32 3)
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR1]])
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB3]], i32 6, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, %struct.S.0*, %struct.S.0*, i32*, [2 x i32]*, [2 x %struct.S.0]*)* @.omp_outlined..1 to void (i32*, i32*, ...)*), i32* [[T_VAR]], %struct.S.0* [[VAR]], %struct.S.0* [[VAR1]], i32* [[T_VAR1]], [2 x i32]* [[VEC]], [2 x %struct.S.0]* [[S_ARR]])
// CHECK2-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR1]]) #[[ATTR4]]
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR]]) #[[ATTR4]]
// CHECK2-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x %struct.S.0], [2 x %struct.S.0]* [[S_ARR]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [[STRUCT_S_0]], %struct.S.0* [[ARRAY_BEGIN]], i64 2
// CHECK2-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK2:       arraydestroy.body:
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S.0* [ [[TMP1]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S_0]], %struct.S.0* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR4]]
// CHECK2-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S.0* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN]]
// CHECK2-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE1:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK2:       arraydestroy.done1:
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[TEST]]) #[[ATTR4]]
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    ret i32 [[TMP2]]
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfEC2Ev
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP0:%.*]] = load volatile double, double* @g, align 8
// CHECK2-NEXT:    [[CONV:%.*]] = fptrunc double [[TMP0]] to float
// CHECK2-NEXT:    store float [[CONV]], float* [[F]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfEC2Ef
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], float [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca float, align 4
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store float [[A]], float* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP0:%.*]] = load float, float* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = fpext float [[TMP0]] to double
// CHECK2-NEXT:    [[TMP1:%.*]] = load volatile double, double* @g, align 8
// CHECK2-NEXT:    [[ADD:%.*]] = fadd double [[CONV]], [[TMP1]]
// CHECK2-NEXT:    [[CONV2:%.*]] = fptrunc double [[ADD]] to float
// CHECK2-NEXT:    store float [[CONV2]], float* [[F]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIfED2Ev
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiEC1Ev
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    call void @_ZN1SIiEC2Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS1]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiEC1Ei
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
// CHECK2-NEXT:    call void @_ZN1SIiEC2Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS1]], i32 [[TMP0]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined..1
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[T_VAR:%.*]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR:%.*]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR1:%.*]], i32* nonnull align 4 dereferenceable(4) [[T_VAR1:%.*]], [2 x i32]* nonnull align 4 dereferenceable(8) [[VEC:%.*]], [2 x %struct.S.0]* nonnull align 4 dereferenceable(8) [[S_ARR:%.*]]) #[[ATTR3]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[T_VAR_ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[VAR_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    [[VAR1_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    [[T_VAR1_ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[VEC_ADDR:%.*]] = alloca [2 x i32]*, align 8
// CHECK2-NEXT:    [[S_ARR_ADDR:%.*]] = alloca [2 x %struct.S.0]*, align 8
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_LB_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_UB_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_ST_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_IL_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_SECTIONS_IV_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[T_VAR2:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[VAR3:%.*]] = alloca [[STRUCT_S_0:%.*]], align 4
// CHECK2-NEXT:    [[VAR14:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK2-NEXT:    [[T_VAR15:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [4 x i8*], align 8
// CHECK2-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK2-NEXT:    [[REF_TMP13:%.*]] = alloca [[STRUCT_S_0]], align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
// CHECK2-NEXT:    store i32* [[T_VAR]], i32** [[T_VAR_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S.0* [[VAR]], %struct.S.0** [[VAR_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S.0* [[VAR1]], %struct.S.0** [[VAR1_ADDR]], align 8
// CHECK2-NEXT:    store i32* [[T_VAR1]], i32** [[T_VAR1_ADDR]], align 8
// CHECK2-NEXT:    store [2 x i32]* [[VEC]], [2 x i32]** [[VEC_ADDR]], align 8
// CHECK2-NEXT:    store [2 x %struct.S.0]* [[S_ARR]], [2 x %struct.S.0]** [[S_ARR_ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[T_VAR_ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load %struct.S.0*, %struct.S.0** [[VAR_ADDR]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load %struct.S.0*, %struct.S.0** [[VAR1_ADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32*, i32** [[T_VAR1_ADDR]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = load [2 x i32]*, [2 x i32]** [[VEC_ADDR]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = load [2 x %struct.S.0]*, [2 x %struct.S.0]** [[S_ARR_ADDR]], align 8
// CHECK2-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK2-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_ST_]], align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_IL_]], align 4
// CHECK2-NEXT:    store i32 0, i32* [[T_VAR2]], align 4
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK2-NEXT:    store i32 2147483647, i32* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4
// CHECK2-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB1]], i32 [[TMP7]], i32 34, i32* [[DOTOMP_SECTIONS_IL_]], i32* [[DOTOMP_SECTIONS_LB_]], i32* [[DOTOMP_SECTIONS_UB_]], i32* [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
// CHECK2-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    [[TMP9:%.*]] = icmp slt i32 [[TMP8]], 1
// CHECK2-NEXT:    [[TMP10:%.*]] = select i1 [[TMP9]], i32 [[TMP8]], i32 1
// CHECK2-NEXT:    store i32 [[TMP10]], i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    [[TMP11:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK2-NEXT:    store i32 [[TMP11]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK2:       omp.inner.for.cond:
// CHECK2-NEXT:    [[TMP12:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    [[TMP13:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK2-NEXT:    [[CMP:%.*]] = icmp sle i32 [[TMP12]], [[TMP13]]
// CHECK2-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK2:       omp.inner.for.body:
// CHECK2-NEXT:    [[TMP14:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    switch i32 [[TMP14]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK2-NEXT:    i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK2-NEXT:    i32 1, label [[DOTOMP_SECTIONS_CASE6:%.*]]
// CHECK2-NEXT:    ]
// CHECK2:       .omp.sections.case:
// CHECK2-NEXT:    [[TMP15:%.*]] = load i32, i32* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], [2 x i32]* [[TMP4]], i64 0, i64 0
// CHECK2-NEXT:    store i32 [[TMP15]], i32* [[ARRAYIDX]], align 4
// CHECK2-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK2:       .omp.sections.case6:
// CHECK2-NEXT:    [[ARRAYIDX7:%.*]] = getelementptr inbounds [2 x %struct.S.0], [2 x %struct.S.0]* [[TMP5]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP16:%.*]] = bitcast %struct.S.0* [[ARRAYIDX7]] to i8*
// CHECK2-NEXT:    [[TMP17:%.*]] = bitcast %struct.S.0* [[VAR3]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP16]], i8* align 4 [[TMP17]], i64 4, i1 false)
// CHECK2-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK2:       .omp.sections.exit:
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK2:       omp.inner.for.inc:
// CHECK2-NEXT:    [[TMP18:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP18]], 1
// CHECK2-NEXT:    store i32 [[INC]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK2:       omp.inner.for.end:
// CHECK2-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[TMP7]])
// CHECK2-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP20:%.*]] = bitcast i32* [[T_VAR2]] to i8*
// CHECK2-NEXT:    store i8* [[TMP20]], i8** [[TMP19]], align 8
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP22:%.*]] = bitcast %struct.S.0* [[VAR3]] to i8*
// CHECK2-NEXT:    store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK2-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP24:%.*]] = bitcast %struct.S.0* [[VAR14]] to i8*
// CHECK2-NEXT:    store i8* [[TMP24]], i8** [[TMP23]], align 8
// CHECK2-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
// CHECK2-NEXT:    [[TMP26:%.*]] = bitcast i32* [[T_VAR15]] to i8*
// CHECK2-NEXT:    store i8* [[TMP26]], i8** [[TMP25]], align 8
// CHECK2-NEXT:    [[TMP27:%.*]] = bitcast [4 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT:    [[TMP28:%.*]] = call i32 @__kmpc_reduce_nowait(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], i32 4, i64 32, i8* [[TMP27]], void (i8*, i8*)* @.omp.reduction.reduction_func.2, [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK2-NEXT:    switch i32 [[TMP28]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK2-NEXT:    i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK2-NEXT:    i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK2-NEXT:    ]
// CHECK2:       .omp.reduction.case1:
// CHECK2-NEXT:    [[TMP29:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[TMP30:%.*]] = load i32, i32* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP29]], [[TMP30]]
// CHECK2-NEXT:    store i32 [[ADD]], i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S.0* @_ZN1SIiEanERKS0_(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK2-NEXT:    [[TMP31:%.*]] = bitcast %struct.S.0* [[TMP1]] to i8*
// CHECK2-NEXT:    [[TMP32:%.*]] = bitcast %struct.S.0* [[CALL]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP31]], i8* align 4 [[TMP32]], i64 4, i1 false)
// CHECK2-NEXT:    [[CALL8:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK2-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[CALL8]], 0
// CHECK2-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK2:       land.rhs:
// CHECK2-NEXT:    [[CALL9:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK2-NEXT:    [[TOBOOL10:%.*]] = icmp ne i32 [[CALL9]], 0
// CHECK2-NEXT:    br label [[LAND_END]]
// CHECK2:       land.end:
// CHECK2-NEXT:    [[TMP33:%.*]] = phi i1 [ false, [[DOTOMP_REDUCTION_CASE1]] ], [ [[TOBOOL10]], [[LAND_RHS]] ]
// CHECK2-NEXT:    [[CONV:%.*]] = zext i1 [[TMP33]] to i32
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]], i32 [[CONV]])
// CHECK2-NEXT:    [[TMP34:%.*]] = bitcast %struct.S.0* [[TMP2]] to i8*
// CHECK2-NEXT:    [[TMP35:%.*]] = bitcast %struct.S.0* [[REF_TMP]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP34]], i8* align 4 [[TMP35]], i64 4, i1 false)
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK2-NEXT:    [[TMP36:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK2-NEXT:    [[TMP37:%.*]] = load i32, i32* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[CMP11:%.*]] = icmp slt i32 [[TMP36]], [[TMP37]]
// CHECK2-NEXT:    br i1 [[CMP11]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK2:       cond.true:
// CHECK2-NEXT:    [[TMP38:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK2-NEXT:    br label [[COND_END:%.*]]
// CHECK2:       cond.false:
// CHECK2-NEXT:    [[TMP39:%.*]] = load i32, i32* [[T_VAR15]], align 4
// CHECK2-NEXT:    br label [[COND_END]]
// CHECK2:       cond.end:
// CHECK2-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP38]], [[COND_TRUE]] ], [ [[TMP39]], [[COND_FALSE]] ]
// CHECK2-NEXT:    store i32 [[COND]], i32* [[TMP3]], align 4
// CHECK2-NEXT:    call void @__kmpc_end_reduce_nowait(%struct.ident_t* @[[GLOB2]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK2-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK2:       .omp.reduction.case2:
// CHECK2-NEXT:    [[TMP40:%.*]] = load i32, i32* [[T_VAR2]], align 4
// CHECK2-NEXT:    [[TMP41:%.*]] = atomicrmw add i32* [[TMP0]], i32 [[TMP40]] monotonic, align 4
// CHECK2-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    [[CALL12:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S.0* @_ZN1SIiEanERKS0_(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP1]], %struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK2-NEXT:    [[TMP42:%.*]] = bitcast %struct.S.0* [[TMP1]] to i8*
// CHECK2-NEXT:    [[TMP43:%.*]] = bitcast %struct.S.0* [[CALL12]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP42]], i8* align 4 [[TMP43]], i64 4, i1 false)
// CHECK2-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    call void @__kmpc_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    [[CALL14:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP2]])
// CHECK2-NEXT:    [[TOBOOL15:%.*]] = icmp ne i32 [[CALL14]], 0
// CHECK2-NEXT:    br i1 [[TOBOOL15]], label [[LAND_RHS16:%.*]], label [[LAND_END19:%.*]]
// CHECK2:       land.rhs16:
// CHECK2-NEXT:    [[CALL17:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]])
// CHECK2-NEXT:    [[TOBOOL18:%.*]] = icmp ne i32 [[CALL17]], 0
// CHECK2-NEXT:    br label [[LAND_END19]]
// CHECK2:       land.end19:
// CHECK2-NEXT:    [[TMP44:%.*]] = phi i1 [ false, [[DOTOMP_REDUCTION_CASE2]] ], [ [[TOBOOL18]], [[LAND_RHS16]] ]
// CHECK2-NEXT:    [[CONV20:%.*]] = zext i1 [[TMP44]] to i32
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP13]], i32 [[CONV20]])
// CHECK2-NEXT:    [[TMP45:%.*]] = bitcast %struct.S.0* [[TMP2]] to i8*
// CHECK2-NEXT:    [[TMP46:%.*]] = bitcast %struct.S.0* [[REF_TMP13]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP45]], i8* align 4 [[TMP46]], i64 4, i1 false)
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP13]]) #[[ATTR4]]
// CHECK2-NEXT:    call void @__kmpc_end_critical(%struct.ident_t* @[[GLOB3]], i32 [[TMP7]], [8 x i32]* @.gomp_critical_user_.atomic_reduction.var)
// CHECK2-NEXT:    [[TMP47:%.*]] = load i32, i32* [[T_VAR15]], align 4
// CHECK2-NEXT:    [[TMP48:%.*]] = atomicrmw min i32* [[TMP3]], i32 [[TMP47]] monotonic, align 4
// CHECK2-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK2:       .omp.reduction.default:
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR14]]) #[[ATTR4]]
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[VAR3]]) #[[ATTR4]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp.reduction.reduction_func.2
// CHECK2-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[REF_TMP:%.*]] = alloca [[STRUCT_S_0:%.*]], align 4
// CHECK2-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP2]] to [4 x i8*]*
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to [4 x i8*]*
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK2-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to i32*
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP10:%.*]] = load i8*, i8** [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to i32*
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[TMP12]], align 8
// CHECK2-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to %struct.S.0*
// CHECK2-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP16:%.*]] = load i8*, i8** [[TMP15]], align 8
// CHECK2-NEXT:    [[TMP17:%.*]] = bitcast i8* [[TMP16]] to %struct.S.0*
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    [[TMP20:%.*]] = bitcast i8* [[TMP19]] to %struct.S.0*
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP22:%.*]] = load i8*, i8** [[TMP21]], align 8
// CHECK2-NEXT:    [[TMP23:%.*]] = bitcast i8* [[TMP22]] to %struct.S.0*
// CHECK2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP5]], i64 0, i64 3
// CHECK2-NEXT:    [[TMP25:%.*]] = load i8*, i8** [[TMP24]], align 8
// CHECK2-NEXT:    [[TMP26:%.*]] = bitcast i8* [[TMP25]] to i32*
// CHECK2-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[TMP3]], i64 0, i64 3
// CHECK2-NEXT:    [[TMP28:%.*]] = load i8*, i8** [[TMP27]], align 8
// CHECK2-NEXT:    [[TMP29:%.*]] = bitcast i8* [[TMP28]] to i32*
// CHECK2-NEXT:    [[TMP30:%.*]] = load i32, i32* [[TMP11]], align 4
// CHECK2-NEXT:    [[TMP31:%.*]] = load i32, i32* [[TMP8]], align 4
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP30]], [[TMP31]]
// CHECK2-NEXT:    store i32 [[ADD]], i32* [[TMP11]], align 4
// CHECK2-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S.0* @_ZN1SIiEanERKS0_(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP17]], %struct.S.0* nonnull align 4 dereferenceable(4) [[TMP14]])
// CHECK2-NEXT:    [[TMP32:%.*]] = bitcast %struct.S.0* [[TMP17]] to i8*
// CHECK2-NEXT:    [[TMP33:%.*]] = bitcast %struct.S.0* [[CALL]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP32]], i8* align 4 [[TMP33]], i64 4, i1 false)
// CHECK2-NEXT:    [[CALL2:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP23]])
// CHECK2-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[CALL2]], 0
// CHECK2-NEXT:    br i1 [[TOBOOL]], label [[LAND_RHS:%.*]], label [[LAND_END:%.*]]
// CHECK2:       land.rhs:
// CHECK2-NEXT:    [[CALL3:%.*]] = call i32 @_ZN1SIiEcviEv(%struct.S.0* nonnull align 4 dereferenceable(4) [[TMP20]])
// CHECK2-NEXT:    [[TOBOOL4:%.*]] = icmp ne i32 [[CALL3]], 0
// CHECK2-NEXT:    br label [[LAND_END]]
// CHECK2:       land.end:
// CHECK2-NEXT:    [[TMP34:%.*]] = phi i1 [ false, [[ENTRY:%.*]] ], [ [[TOBOOL4]], [[LAND_RHS]] ]
// CHECK2-NEXT:    [[CONV:%.*]] = zext i1 [[TMP34]] to i32
// CHECK2-NEXT:    call void @_ZN1SIiEC1Ei(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]], i32 [[CONV]])
// CHECK2-NEXT:    [[TMP35:%.*]] = bitcast %struct.S.0* [[TMP23]] to i8*
// CHECK2-NEXT:    [[TMP36:%.*]] = bitcast %struct.S.0* [[REF_TMP]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 [[TMP35]], i8* align 4 [[TMP36]], i64 4, i1 false)
// CHECK2-NEXT:    call void @_ZN1SIiED1Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[REF_TMP]]) #[[ATTR4]]
// CHECK2-NEXT:    [[TMP37:%.*]] = load i32, i32* [[TMP29]], align 4
// CHECK2-NEXT:    [[TMP38:%.*]] = load i32, i32* [[TMP26]], align 4
// CHECK2-NEXT:    [[CMP:%.*]] = icmp slt i32 [[TMP37]], [[TMP38]]
// CHECK2-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK2:       cond.true:
// CHECK2-NEXT:    [[TMP39:%.*]] = load i32, i32* [[TMP29]], align 4
// CHECK2-NEXT:    br label [[COND_END:%.*]]
// CHECK2:       cond.false:
// CHECK2-NEXT:    [[TMP40:%.*]] = load i32, i32* [[TMP26]], align 4
// CHECK2-NEXT:    br label [[COND_END]]
// CHECK2:       cond.end:
// CHECK2-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP39]], [[COND_TRUE]] ], [ [[TMP40]], [[COND_FALSE]] ]
// CHECK2-NEXT:    store i32 [[COND]], i32* [[TMP29]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiEanERKS0_
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]], %struct.S.0* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR6]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S.0* [[TMP0]], %struct.S.0** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    ret %struct.S.0* [[THIS1]]
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiEcviEv
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) #[[ATTR6]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiED1Ev
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    call void @_ZN1SIiED2Ev(%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS1]]) #[[ATTR4]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiEC2Ev
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S_0:%.*]], %struct.S.0* [[THIS1]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP0:%.*]] = load volatile double, double* @g, align 8
// CHECK2-NEXT:    [[CONV:%.*]] = fptosi double [[TMP0]] to i32
// CHECK2-NEXT:    store i32 [[CONV]], i32* [[F]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiEC2Ei
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[A:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[F:%.*]] = getelementptr inbounds [[STRUCT_S_0:%.*]], %struct.S.0* [[THIS1]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = sitofp i32 [[TMP0]] to double
// CHECK2-NEXT:    [[TMP1:%.*]] = load volatile double, double* @g, align 8
// CHECK2-NEXT:    [[ADD:%.*]] = fadd double [[CONV]], [[TMP1]]
// CHECK2-NEXT:    [[CONV2:%.*]] = fptosi double [[ADD]] to i32
// CHECK2-NEXT:    store i32 [[CONV2]], i32* [[F]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SIiED2Ev
// CHECK2-SAME: (%struct.S.0* nonnull align 4 dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S.0*, align 8
// CHECK2-NEXT:    store %struct.S.0* [[THIS]], %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S.0*, %struct.S.0** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@main
// CHECK3-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[REF_TMP:%.*]] = alloca [[CLASS_ANON:%.*]], align 1
// CHECK3-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK3-NEXT:    call void @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 1 dereferenceable(1) [[REF_TMP]])
// CHECK3-NEXT:    ret i32 0
//
//
// CHECK3-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]]) #[[ATTR2:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK3-NEXT:    [[DOTOMP_SECTIONS_LB_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTOMP_SECTIONS_UB_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTOMP_SECTIONS_ST_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTOMP_SECTIONS_IL_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTOMP_SECTIONS_IV_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[G:%.*]] = alloca double, align 8
// CHECK3-NEXT:    [[REF_TMP:%.*]] = alloca [[CLASS_ANON_0:%.*]], align 8
// CHECK3-NEXT:    [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x i8*], align 8
// CHECK3-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca double, align 8
// CHECK3-NEXT:    [[TMP:%.*]] = alloca double, align 8
// CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
// CHECK3-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK3-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK3-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_ST_]], align 4
// CHECK3-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_IL_]], align 4
// CHECK3-NEXT:    store double 0.000000e+00, double* [[G]], align 8
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK3-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK3-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB1:[0-9]+]], i32 [[TMP1]], i32 34, i32* [[DOTOMP_SECTIONS_IL_]], i32* [[DOTOMP_SECTIONS_LB_]], i32* [[DOTOMP_SECTIONS_UB_]], i32* [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
// CHECK3-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK3-NEXT:    [[TMP3:%.*]] = icmp slt i32 [[TMP2]], 1
// CHECK3-NEXT:    [[TMP4:%.*]] = select i1 [[TMP3]], i32 [[TMP2]], i32 1
// CHECK3-NEXT:    store i32 [[TMP4]], i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK3-NEXT:    [[TMP5:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK3-NEXT:    store i32 [[TMP5]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK3-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3:       omp.inner.for.cond:
// CHECK3-NEXT:    [[TMP6:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK3-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK3-NEXT:    [[CMP:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK3-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK3:       omp.inner.for.body:
// CHECK3-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK3-NEXT:    switch i32 [[TMP8]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK3-NEXT:    i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK3-NEXT:    i32 1, label [[DOTOMP_SECTIONS_CASE1:%.*]]
// CHECK3-NEXT:    ]
// CHECK3:       .omp.sections.case:
// CHECK3-NEXT:    store double 1.000000e+00, double* [[G]], align 8
// CHECK3-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK3:       .omp.sections.case1:
// CHECK3-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[REF_TMP]], i32 0, i32 0
// CHECK3-NEXT:    store double* [[G]], double** [[TMP9]], align 8
// CHECK3-NEXT:    call void @"_ZZZ4mainENK3$_0clEvENKUlvE_clEv"(%class.anon.0* nonnull align 8 dereferenceable(8) [[REF_TMP]])
// CHECK3-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK3:       .omp.sections.exit:
// CHECK3-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK3:       omp.inner.for.inc:
// CHECK3-NEXT:    [[TMP10:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK3-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP10]], 1
// CHECK3-NEXT:    store i32 [[INC]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK3-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK3:       omp.inner.for.end:
// CHECK3-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK3-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK3-NEXT:    [[TMP12:%.*]] = bitcast double* [[G]] to i8*
// CHECK3-NEXT:    store i8* [[TMP12]], i8** [[TMP11]], align 8
// CHECK3-NEXT:    [[TMP13:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT:    [[TMP14:%.*]] = call i32 @__kmpc_reduce(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 1, i64 8, i8* [[TMP13]], void (i8*, i8*)* @.omp.reduction.reduction_func, [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT:    switch i32 [[TMP14]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK3-NEXT:    i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK3-NEXT:    i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK3-NEXT:    ]
// CHECK3:       .omp.reduction.case1:
// CHECK3-NEXT:    [[TMP15:%.*]] = load double, double* @g, align 8
// CHECK3-NEXT:    [[TMP16:%.*]] = load double, double* [[G]], align 8
// CHECK3-NEXT:    [[ADD:%.*]] = fadd double [[TMP15]], [[TMP16]]
// CHECK3-NEXT:    store double [[ADD]], double* @g, align 8
// CHECK3-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3:       .omp.reduction.case2:
// CHECK3-NEXT:    [[TMP17:%.*]] = load double, double* [[G]], align 8
// CHECK3-NEXT:    [[ATOMIC_LOAD:%.*]] = load atomic i64, i64* bitcast (double* @g to i64*) monotonic, align 8
// CHECK3-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK3:       atomic_cont:
// CHECK3-NEXT:    [[TMP18:%.*]] = phi i64 [ [[ATOMIC_LOAD]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[TMP25:%.*]], [[ATOMIC_CONT]] ]
// CHECK3-NEXT:    [[TMP19:%.*]] = bitcast double* [[ATOMIC_TEMP]] to i64*
// CHECK3-NEXT:    [[TMP20:%.*]] = bitcast i64 [[TMP18]] to double
// CHECK3-NEXT:    store double [[TMP20]], double* [[TMP]], align 8
// CHECK3-NEXT:    [[TMP21:%.*]] = load double, double* [[TMP]], align 8
// CHECK3-NEXT:    [[TMP22:%.*]] = load double, double* [[G]], align 8
// CHECK3-NEXT:    [[ADD2:%.*]] = fadd double [[TMP21]], [[TMP22]]
// CHECK3-NEXT:    store double [[ADD2]], double* [[ATOMIC_TEMP]], align 8
// CHECK3-NEXT:    [[TMP23:%.*]] = load i64, i64* [[TMP19]], align 8
// CHECK3-NEXT:    [[TMP24:%.*]] = cmpxchg i64* bitcast (double* @g to i64*), i64 [[TMP18]], i64 [[TMP23]] monotonic monotonic, align 8
// CHECK3-NEXT:    [[TMP25]] = extractvalue { i64, i1 } [[TMP24]], 0
// CHECK3-NEXT:    [[TMP26:%.*]] = extractvalue { i64, i1 } [[TMP24]], 1
// CHECK3-NEXT:    br i1 [[TMP26]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK3:       atomic_exit:
// CHECK3-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3:       .omp.reduction.default:
// CHECK3-NEXT:    call void @__kmpc_barrier(%struct.ident_t* @[[GLOB3:[0-9]+]], i32 [[TMP1]])
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@.omp.reduction.reduction_func
// CHECK3-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK3-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK3-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK3-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK3-NEXT:    [[TMP2:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK3-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP2]] to [1 x i8*]*
// CHECK3-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK3-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to [1 x i8*]*
// CHECK3-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[TMP5]], i64 0, i64 0
// CHECK3-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK3-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to double*
// CHECK3-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[TMP3]], i64 0, i64 0
// CHECK3-NEXT:    [[TMP10:%.*]] = load i8*, i8** [[TMP9]], align 8
// CHECK3-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to double*
// CHECK3-NEXT:    [[TMP12:%.*]] = load double, double* [[TMP11]], align 8
// CHECK3-NEXT:    [[TMP13:%.*]] = load double, double* [[TMP8]], align 8
// CHECK3-NEXT:    [[ADD:%.*]] = fadd double [[TMP12]], [[TMP13]]
// CHECK3-NEXT:    store double [[ADD]], double* [[TMP11]], align 8
// CHECK3-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@main
// CHECK4-SAME: () #[[ATTR1:[0-9]+]] {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK4-NEXT:    [[TMP0:%.*]] = load i8*, i8** getelementptr inbounds ([[STRUCT___BLOCK_LITERAL_GENERIC:%.*]], %struct.__block_literal_generic* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor* }* @__block_literal_global to %struct.__block_literal_generic*), i32 0, i32 3), align 8
// CHECK4-NEXT:    [[TMP1:%.*]] = bitcast i8* [[TMP0]] to void (i8*)*
// CHECK4-NEXT:    call void [[TMP1]](i8* bitcast ({ i8**, i32, i32, i8*, %struct.__block_descriptor* }* @__block_literal_global to i8*))
// CHECK4-NEXT:    ret i32 0
//
//
// CHECK4-LABEL: define {{[^@]+}}@__main_block_invoke
// CHECK4-SAME: (i8* [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2:[0-9]+]] {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    [[BLOCK_ADDR:%.*]] = alloca <{ i8*, i32, i32, i8*, %struct.__block_descriptor* }>*, align 8
// CHECK4-NEXT:    store i8* [[DOTBLOCK_DESCRIPTOR]], i8** [[DOTBLOCK_DESCRIPTOR_ADDR]], align 8
// CHECK4-NEXT:    [[BLOCK:%.*]] = bitcast i8* [[DOTBLOCK_DESCRIPTOR]] to <{ i8*, i32, i32, i8*, %struct.__block_descriptor* }>*
// CHECK4-NEXT:    store <{ i8*, i32, i32, i8*, %struct.__block_descriptor* }>* [[BLOCK]], <{ i8*, i32, i32, i8*, %struct.__block_descriptor* }>** [[BLOCK_ADDR]], align 8
// CHECK4-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB4:[0-9]+]], i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* @.omp_outlined. to void (i32*, i32*, ...)*))
// CHECK4-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK4-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK4-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK4-NEXT:    [[DOTOMP_SECTIONS_LB_:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[DOTOMP_SECTIONS_UB_:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[DOTOMP_SECTIONS_ST_:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[DOTOMP_SECTIONS_IL_:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[DOTOMP_SECTIONS_IV_:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[G:%.*]] = alloca double, align 8
// CHECK4-NEXT:    [[BLOCK:%.*]] = alloca <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, align 8
// CHECK4-NEXT:    [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x i8*], align 8
// CHECK4-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca double, align 8
// CHECK4-NEXT:    [[TMP:%.*]] = alloca double, align 8
// CHECK4-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK4-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
// CHECK4-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK4-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK4-NEXT:    store i32 1, i32* [[DOTOMP_SECTIONS_ST_]], align 4
// CHECK4-NEXT:    store i32 0, i32* [[DOTOMP_SECTIONS_IL_]], align 4
// CHECK4-NEXT:    store double 0.000000e+00, double* [[G]], align 8
// CHECK4-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK4-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK4-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB1:[0-9]+]], i32 [[TMP1]], i32 34, i32* [[DOTOMP_SECTIONS_IL_]], i32* [[DOTOMP_SECTIONS_LB_]], i32* [[DOTOMP_SECTIONS_UB_]], i32* [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
// CHECK4-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK4-NEXT:    [[TMP3:%.*]] = icmp slt i32 [[TMP2]], 1
// CHECK4-NEXT:    [[TMP4:%.*]] = select i1 [[TMP3]], i32 [[TMP2]], i32 1
// CHECK4-NEXT:    store i32 [[TMP4]], i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK4-NEXT:    [[TMP5:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK4-NEXT:    store i32 [[TMP5]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK4-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK4:       omp.inner.for.cond:
// CHECK4-NEXT:    [[TMP6:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK4-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK4-NEXT:    [[CMP:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK4-NEXT:    br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK4:       omp.inner.for.body:
// CHECK4-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK4-NEXT:    switch i32 [[TMP8]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK4-NEXT:    i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK4-NEXT:    i32 1, label [[DOTOMP_SECTIONS_CASE1:%.*]]
// CHECK4-NEXT:    ]
// CHECK4:       .omp.sections.case:
// CHECK4-NEXT:    store double 1.000000e+00, double* [[G]], align 8
// CHECK4-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK4:       .omp.sections.case1:
// CHECK4-NEXT:    [[BLOCK_ISA:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 0
// CHECK4-NEXT:    store i8* bitcast (i8** @_NSConcreteStackBlock to i8*), i8** [[BLOCK_ISA]], align 8
// CHECK4-NEXT:    [[BLOCK_FLAGS:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 1
// CHECK4-NEXT:    store i32 1073741824, i32* [[BLOCK_FLAGS]], align 8
// CHECK4-NEXT:    [[BLOCK_RESERVED:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 2
// CHECK4-NEXT:    store i32 0, i32* [[BLOCK_RESERVED]], align 4
// CHECK4-NEXT:    [[BLOCK_INVOKE:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 3
// CHECK4-NEXT:    store i8* bitcast (void (i8*)* @_block_invoke to i8*), i8** [[BLOCK_INVOKE]], align 8
// CHECK4-NEXT:    [[BLOCK_DESCRIPTOR:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 4
// CHECK4-NEXT:    store %struct.__block_descriptor* bitcast ({ i64, i64, i8*, i8* }* @__block_descriptor_tmp.1 to %struct.__block_descriptor*), %struct.__block_descriptor** [[BLOCK_DESCRIPTOR]], align 8
// CHECK4-NEXT:    [[BLOCK_CAPTURED:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 5
// CHECK4-NEXT:    [[TMP9:%.*]] = load volatile double, double* [[G]], align 8
// CHECK4-NEXT:    store volatile double [[TMP9]], double* [[BLOCK_CAPTURED]], align 8
// CHECK4-NEXT:    [[TMP10:%.*]] = bitcast <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]] to void ()*
// CHECK4-NEXT:    [[BLOCK_LITERAL:%.*]] = bitcast void ()* [[TMP10]] to %struct.__block_literal_generic*
// CHECK4-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___BLOCK_LITERAL_GENERIC:%.*]], %struct.__block_literal_generic* [[BLOCK_LITERAL]], i32 0, i32 3
// CHECK4-NEXT:    [[TMP12:%.*]] = bitcast %struct.__block_literal_generic* [[BLOCK_LITERAL]] to i8*
// CHECK4-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[TMP11]], align 8
// CHECK4-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to void (i8*)*
// CHECK4-NEXT:    call void [[TMP14]](i8* [[TMP12]])
// CHECK4-NEXT:    br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK4:       .omp.sections.exit:
// CHECK4-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK4:       omp.inner.for.inc:
// CHECK4-NEXT:    [[TMP15:%.*]] = load i32, i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK4-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP15]], 1
// CHECK4-NEXT:    store i32 [[INC]], i32* [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK4-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK4:       omp.inner.for.end:
// CHECK4-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK4-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK4-NEXT:    [[TMP17:%.*]] = bitcast double* [[G]] to i8*
// CHECK4-NEXT:    store i8* [[TMP17]], i8** [[TMP16]], align 8
// CHECK4-NEXT:    [[TMP18:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK4-NEXT:    [[TMP19:%.*]] = call i32 @__kmpc_reduce(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 1, i64 8, i8* [[TMP18]], void (i8*, i8*)* @.omp.reduction.reduction_func, [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK4-NEXT:    switch i32 [[TMP19]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK4-NEXT:    i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK4-NEXT:    i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK4-NEXT:    ]
// CHECK4:       .omp.reduction.case1:
// CHECK4-NEXT:    [[TMP20:%.*]] = load double, double* @g, align 8
// CHECK4-NEXT:    [[TMP21:%.*]] = load double, double* [[G]], align 8
// CHECK4-NEXT:    [[ADD:%.*]] = fadd double [[TMP20]], [[TMP21]]
// CHECK4-NEXT:    store double [[ADD]], double* @g, align 8
// CHECK4-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK4-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK4:       .omp.reduction.case2:
// CHECK4-NEXT:    [[TMP22:%.*]] = load double, double* [[G]], align 8
// CHECK4-NEXT:    [[ATOMIC_LOAD:%.*]] = load atomic i64, i64* bitcast (double* @g to i64*) monotonic, align 8
// CHECK4-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK4:       atomic_cont:
// CHECK4-NEXT:    [[TMP23:%.*]] = phi i64 [ [[ATOMIC_LOAD]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[TMP30:%.*]], [[ATOMIC_CONT]] ]
// CHECK4-NEXT:    [[TMP24:%.*]] = bitcast double* [[ATOMIC_TEMP]] to i64*
// CHECK4-NEXT:    [[TMP25:%.*]] = bitcast i64 [[TMP23]] to double
// CHECK4-NEXT:    store double [[TMP25]], double* [[TMP]], align 8
// CHECK4-NEXT:    [[TMP26:%.*]] = load double, double* [[TMP]], align 8
// CHECK4-NEXT:    [[TMP27:%.*]] = load double, double* [[G]], align 8
// CHECK4-NEXT:    [[ADD2:%.*]] = fadd double [[TMP26]], [[TMP27]]
// CHECK4-NEXT:    store double [[ADD2]], double* [[ATOMIC_TEMP]], align 8
// CHECK4-NEXT:    [[TMP28:%.*]] = load i64, i64* [[TMP24]], align 8
// CHECK4-NEXT:    [[TMP29:%.*]] = cmpxchg i64* bitcast (double* @g to i64*), i64 [[TMP23]], i64 [[TMP28]] monotonic monotonic, align 8
// CHECK4-NEXT:    [[TMP30]] = extractvalue { i64, i1 } [[TMP29]], 0
// CHECK4-NEXT:    [[TMP31:%.*]] = extractvalue { i64, i1 } [[TMP29]], 1
// CHECK4-NEXT:    br i1 [[TMP31]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK4:       atomic_exit:
// CHECK4-NEXT:    call void @__kmpc_end_reduce(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], [8 x i32]* @.gomp_critical_user_.reduction.var)
// CHECK4-NEXT:    br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK4:       .omp.reduction.default:
// CHECK4-NEXT:    call void @__kmpc_barrier(%struct.ident_t* @[[GLOB3:[0-9]+]], i32 [[TMP1]])
// CHECK4-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@_block_invoke
// CHECK4-SAME: (i8* [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2]] {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    [[BLOCK_ADDR:%.*]] = alloca <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>*, align 8
// CHECK4-NEXT:    store i8* [[DOTBLOCK_DESCRIPTOR]], i8** [[DOTBLOCK_DESCRIPTOR_ADDR]], align 8
// CHECK4-NEXT:    [[BLOCK:%.*]] = bitcast i8* [[DOTBLOCK_DESCRIPTOR]] to <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>*
// CHECK4-NEXT:    store <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>** [[BLOCK_ADDR]], align 8
// CHECK4-NEXT:    [[BLOCK_CAPTURE_ADDR:%.*]] = getelementptr inbounds <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>, <{ i8*, i32, i32, i8*, %struct.__block_descriptor*, double }>* [[BLOCK]], i32 0, i32 5
// CHECK4-NEXT:    store double 2.000000e+00, double* [[BLOCK_CAPTURE_ADDR]], align 8
// CHECK4-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@.omp.reduction.reduction_func
// CHECK4-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK4-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK4-NEXT:    [[TMP2:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK4-NEXT:    [[TMP3:%.*]] = bitcast i8* [[TMP2]] to [1 x i8*]*
// CHECK4-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK4-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to [1 x i8*]*
// CHECK4-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[TMP5]], i64 0, i64 0
// CHECK4-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK4-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to double*
// CHECK4-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[TMP3]], i64 0, i64 0
// CHECK4-NEXT:    [[TMP10:%.*]] = load i8*, i8** [[TMP9]], align 8
// CHECK4-NEXT:    [[TMP11:%.*]] = bitcast i8* [[TMP10]] to double*
// CHECK4-NEXT:    [[TMP12:%.*]] = load double, double* [[TMP11]], align 8
// CHECK4-NEXT:    [[TMP13:%.*]] = load double, double* [[TMP8]], align 8
// CHECK4-NEXT:    [[ADD:%.*]] = fadd double [[TMP12]], [[TMP13]]
// CHECK4-NEXT:    store double [[ADD]], double* [[TMP11]], align 8
// CHECK4-NEXT:    ret void
//
//