xref: /llvm-project-15.0.7/clang/test/OpenMP/master_taskloop_in_reduction_codegen.cpp (revision 207b08a9)

// 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 -triple x86_64-apple-darwin10 -fopenmp -x c++ -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK2

// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp-simd -x c++ -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK3
// RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
// RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK4
// expected-no-diagnostics
#ifndef HEADER
#define HEADER


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


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



#endif
// CHECK1-LABEL: define {{[^@]+}}@main
// CHECK1-SAME: (i32 [[ARGC:%.*]], i8** [[ARGV:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[ARGV_ADDR:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    [[A:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[B:%.*]] = alloca float, align 4
// CHECK1-NEXT:    [[C:%.*]] = alloca [5 x %struct.S], align 16
// CHECK1-NEXT:    [[SAVED_STACK:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTRD_INPUT_:%.*]] = alloca [3 x %struct.kmp_taskred_input_t], align 8
// CHECK1-NEXT:    [[DOTTASK_RED_:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTRD_INPUT_3:%.*]] = alloca [2 x %struct.kmp_taskred_input_t.0], align 8
// CHECK1-NEXT:    [[DOTTASK_RED_6:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1:[0-9]+]])
// CHECK1-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    store i32 [[ARGC]], i32* [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i8** [[ARGV]], i8*** [[ARGV_ADDR]], align 8
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK1-NEXT:    [[ARRAYCTOR_END:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK1-NEXT:    br label [[ARRAYCTOR_LOOP:%.*]]
// CHECK1:       arrayctor.loop:
// CHECK1-NEXT:    [[ARRAYCTOR_CUR:%.*]] = phi %struct.S* [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[ARRAYCTOR_NEXT:%.*]], [[ARRAYCTOR_LOOP]] ]
// CHECK1-NEXT:    call void @_ZN1SC1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYCTOR_CUR]])
// CHECK1-NEXT:    [[ARRAYCTOR_NEXT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYCTOR_CUR]], i64 1
// CHECK1-NEXT:    [[ARRAYCTOR_DONE:%.*]] = icmp eq %struct.S* [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK1-NEXT:    br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK1:       arrayctor.cont:
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = zext i32 [[TMP1]] to i64
// CHECK1-NEXT:    [[TMP3:%.*]] = call i8* @llvm.stacksave()
// CHECK1-NEXT:    store i8* [[TMP3]], i8** [[SAVED_STACK]], align 8
// CHECK1-NEXT:    [[VLA:%.*]] = alloca i16, i64 [[TMP2]], align 16
// CHECK1-NEXT:    store i64 [[TMP2]], i64* [[__VLA_EXPR0]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T:%.*]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i32* [[A]] to i8*
// CHECK1-NEXT:    store i8* [[TMP5]], i8** [[TMP4]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP7:%.*]] = bitcast i32* [[A]] to i8*
// CHECK1-NEXT:    store i8* [[TMP7]], i8** [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 2
// CHECK1-NEXT:    store i64 4, i64* [[TMP8]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 3
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init. to i8*), i8** [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 4
// CHECK1-NEXT:    store i8* null, i8** [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 5
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb. to i8*), i8** [[TMP11]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP13:%.*]] = bitcast i32* [[TMP12]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP13]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_1:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP15:%.*]] = bitcast float* [[B]] to i8*
// CHECK1-NEXT:    store i8* [[TMP15]], i8** [[TMP14]], align 8
// CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP17:%.*]] = bitcast float* [[B]] to i8*
// CHECK1-NEXT:    store i8* [[TMP17]], i8** [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 2
// CHECK1-NEXT:    store i64 4, i64* [[TMP18]], align 8
// CHECK1-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 3
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..1 to i8*), i8** [[TMP19]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 4
// CHECK1-NEXT:    store i8* null, i8** [[TMP20]], align 8
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 5
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..2 to i8*), i8** [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP23:%.*]] = bitcast i32* [[TMP22]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP23]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_2:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP25:%.*]] = bitcast i32* [[ARGC_ADDR]] to i8*
// CHECK1-NEXT:    store i8* [[TMP25]], i8** [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP27:%.*]] = bitcast i32* [[ARGC_ADDR]] to i8*
// CHECK1-NEXT:    store i8* [[TMP27]], i8** [[TMP26]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 2
// CHECK1-NEXT:    store i64 4, i64* [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP29:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 3
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..3 to i8*), i8** [[TMP29]], align 8
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 4
// CHECK1-NEXT:    store i8* null, i8** [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP31:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 5
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..4 to i8*), i8** [[TMP31]], align 8
// CHECK1-NEXT:    [[TMP32:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP33]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[TMP34:%.*]] = bitcast [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]] to i8*
// CHECK1-NEXT:    [[TMP35:%.*]] = call i8* @__kmpc_taskred_init(i32 [[TMP0]], i32 3, i8* [[TMP34]])
// CHECK1-NEXT:    store i8* [[TMP35]], i8** [[DOTTASK_RED_]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_4:%.*]] = getelementptr inbounds [2 x %struct.kmp_taskred_input_t.0], [2 x %struct.kmp_taskred_input_t.0]* [[DOTRD_INPUT_3]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0:%.*]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP37:%.*]] = bitcast [5 x %struct.S]* [[C]] to i8*
// CHECK1-NEXT:    store i8* [[TMP37]], i8** [[TMP36]], align 8
// CHECK1-NEXT:    [[TMP38:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP39:%.*]] = bitcast [5 x %struct.S]* [[C]] to i8*
// CHECK1-NEXT:    store i8* [[TMP39]], i8** [[TMP38]], align 8
// CHECK1-NEXT:    [[TMP40:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK1-NEXT:    store i64 20, i64* [[TMP40]], align 8
// CHECK1-NEXT:    [[TMP41:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..5 to i8*), i8** [[TMP41]], align 8
// CHECK1-NEXT:    [[TMP42:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 4
// CHECK1-NEXT:    store i8* bitcast (void (i8*)* @.red_fini. to i8*), i8** [[TMP42]], align 8
// CHECK1-NEXT:    [[TMP43:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 5
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..6 to i8*), i8** [[TMP43]], align 8
// CHECK1-NEXT:    [[TMP44:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP45:%.*]] = bitcast i32* [[TMP44]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP45]], i8 0, i64 4, i1 false)
// CHECK1-NEXT:    [[DOTRD_INPUT_GEP_5:%.*]] = getelementptr inbounds [2 x %struct.kmp_taskred_input_t.0], [2 x %struct.kmp_taskred_input_t.0]* [[DOTRD_INPUT_3]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP46:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP47:%.*]] = bitcast i16* [[VLA]] to i8*
// CHECK1-NEXT:    store i8* [[TMP47]], i8** [[TMP46]], align 8
// CHECK1-NEXT:    [[TMP48:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP49:%.*]] = bitcast i16* [[VLA]] to i8*
// CHECK1-NEXT:    store i8* [[TMP49]], i8** [[TMP48]], align 8
// CHECK1-NEXT:    [[TMP50:%.*]] = mul nuw i64 [[TMP2]], 2
// CHECK1-NEXT:    [[TMP51:%.*]] = udiv exact i64 [[TMP50]], ptrtoint (i16* getelementptr (i16, i16* null, i32 1) to i64)
// CHECK1-NEXT:    [[TMP52:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT:    store i64 [[TMP50]], i64* [[TMP52]], align 8
// CHECK1-NEXT:    [[TMP53:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..7 to i8*), i8** [[TMP53]], align 8
// CHECK1-NEXT:    [[TMP54:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 4
// CHECK1-NEXT:    store i8* null, i8** [[TMP54]], align 8
// CHECK1-NEXT:    [[TMP55:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 5
// CHECK1-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..8 to i8*), i8** [[TMP55]], align 8
// CHECK1-NEXT:    [[TMP56:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 6
// CHECK1-NEXT:    store i32 1, i32* [[TMP56]], align 8
// CHECK1-NEXT:    [[TMP57:%.*]] = bitcast [2 x %struct.kmp_taskred_input_t.0]* [[DOTRD_INPUT_3]] to i8*
// CHECK1-NEXT:    [[TMP58:%.*]] = call i8* @__kmpc_taskred_init(i32 [[TMP0]], i32 2, i8* [[TMP57]])
// CHECK1-NEXT:    store i8* [[TMP58]], i8** [[DOTTASK_RED_6]], align 8
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i16*, i8**, i8**)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* [[A]], i64 [[TMP2]], i16* [[VLA]], i8** [[DOTTASK_RED_]], i8** [[DOTTASK_RED_6]])
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    [[TMP59:%.*]] = load i8*, i8** [[SAVED_STACK]], align 8
// CHECK1-NEXT:    call void @llvm.stackrestore(i8* [[TMP59]])
// CHECK1-NEXT:    [[ARRAY_BEGIN7:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP60:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAY_BEGIN7]], i64 5
// CHECK1-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1:       arraydestroy.body:
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP60]], [[ARRAYCTOR_CONT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3:[0-9]+]]
// CHECK1-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN7]]
// CHECK1-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE8:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK1:       arraydestroy.done8:
// CHECK1-NEXT:    [[TMP61:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    ret i32 [[TMP61]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC1Ev
// CHECK1-SAME: (%struct.S* nonnull 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 @_ZN1SC2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init.
// CHECK1-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 i32*
// CHECK1-NEXT:    store i32 0, i32* [[TMP3]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb.
// 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:    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 i32*
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to i32*
// CHECK1-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP3]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP5]], align 8
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP6]], [[TMP7]]
// CHECK1-NEXT:    store i32 [[ADD]], i32* [[TMP3]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..1
// CHECK1-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 float*
// CHECK1-NEXT:    store float 0.000000e+00, float* [[TMP3]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..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:    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 float*
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to float*
// CHECK1-NEXT:    [[TMP6:%.*]] = load float, float* [[TMP3]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = load float, float* [[TMP5]], align 8
// CHECK1-NEXT:    [[ADD:%.*]] = fadd float [[TMP6]], [[TMP7]]
// CHECK1-NEXT:    store float [[ADD]], float* [[TMP3]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..3
// CHECK1-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 i32*
// CHECK1-NEXT:    store i32 0, i32* [[TMP3]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..4
// 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:    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 i32*
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to i32*
// CHECK1-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP3]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP5]], align 8
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP6]], [[TMP7]]
// CHECK1-NEXT:    store i32 [[ADD]], i32* [[TMP3]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_init..5
// CHECK1-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 [5 x %struct.S]*
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[TMP3]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK1-NEXT:    [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq %struct.S* [[ARRAY_BEGIN]], [[TMP4]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1:       omp.arrayinit.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi %struct.S* [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT:    call void @_ZN1SC1Ev(%struct.S* nonnull dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]])
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], %struct.S* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq %struct.S* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP4]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1:       omp.arrayinit.done:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_fini.
// CHECK1-SAME: (i8* [[TMP0:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = bitcast i8* [[TMP1]] to [5 x %struct.S]*
// CHECK1-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[TMP2]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK1-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1:       arraydestroy.body:
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP3]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]]
// 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:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SD1Ev
// CHECK1-SAME: (%struct.S* nonnull 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 @_ZN1SD2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]]) #[[ATTR3]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..6
// 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:%.*]], 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 %struct.S*
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to %struct.S*
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr [[STRUCT_S]], %struct.S* [[TMP3]], i64 5
// CHECK1-NEXT:    [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq %struct.S* [[TMP3]], [[TMP6]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE2:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1:       omp.arraycpy.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP5]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP3]], [[ENTRY]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    call void @_ZplRK1SS1_(%struct.S* sret([[STRUCT_S]]) align 4 [[REF_TMP]], %struct.S* nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]], %struct.S* nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_SRCELEMENTPAST]])
// CHECK1-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SaSERKS_(%struct.S* nonnull dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]], %struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]])
// CHECK1-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[REF_TMP]]) #[[ATTR3]]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], %struct.S* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr [[STRUCT_S]], %struct.S* [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq %struct.S* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP6]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYCPY_DONE2]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1:       omp.arraycpy.done2:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZplRK1SS1_
// CHECK1-SAME: (%struct.S* noalias sret([[STRUCT_S:%.*]]) align 4 [[AGG_RESULT:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[A:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[B:%.*]]) #[[ATTR7:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[RESULT_PTR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[B_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = bitcast %struct.S* [[AGG_RESULT]] to i8*
// CHECK1-NEXT:    store i8* [[TMP0]], i8** [[RESULT_PTR]], align 8
// CHECK1-NEXT:    store %struct.S* [[A]], %struct.S** [[A_ADDR]], align 8
// CHECK1-NEXT:    store %struct.S* [[B]], %struct.S** [[B_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load %struct.S*, %struct.S** [[A_ADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SC1ERKS_(%struct.S* nonnull dereferenceable(4) [[AGG_RESULT]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SaSERKS_
// CHECK1-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR7]] 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 {{[^@]+}}@.red_init..7
// CHECK1-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i8* bitcast (i64* @{{reduction_size[.].+[.]}})
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to i64*
// CHECK1-NEXT:    [[TMP6:%.*]] = load i64, i64* [[TMP5]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = bitcast i8* [[TMP3]] to i16*
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr i16, i16* [[TMP7]], i64 [[TMP6]]
// CHECK1-NEXT:    [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq i16* [[TMP7]], [[TMP8]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1:       omp.arrayinit.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi i16* [ [[TMP7]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT:    store i16 0, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq i16* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP8]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1:       omp.arrayinit.done:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.red_comb..8
// CHECK1-SAME: (i8* [[TMP0:%.*]], i8* [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i8* bitcast (i64* @{{reduction_size[.].+[.]}})
// CHECK1-NEXT:    [[TMP4:%.*]] = bitcast i8* [[TMP3]] to i64*
// CHECK1-NEXT:    [[TMP5:%.*]] = load i64, i64* [[TMP4]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = bitcast i8* [[TMP6]] to i16*
// CHECK1-NEXT:    [[TMP8:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = bitcast i8* [[TMP8]] to i16*
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr i16, i16* [[TMP7]], i64 [[TMP5]]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq i16* [[TMP7]], [[TMP10]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE4:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1:       omp.arraycpy.body:
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi i16* [ [[TMP9]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi i16* [ [[TMP7]], [[ENTRY]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT:    [[TMP11:%.*]] = load i16, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT:    [[CONV:%.*]] = sext i16 [[TMP11]] to i32
// CHECK1-NEXT:    [[TMP12:%.*]] = load i16, i16* [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
// CHECK1-NEXT:    [[CONV2:%.*]] = sext i16 [[TMP12]] to i32
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], [[CONV2]]
// CHECK1-NEXT:    [[CONV3:%.*]] = trunc i32 [[ADD]] to i16
// CHECK1-NEXT:    store i16 [[CONV3]], i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i16, i16* [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK1-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq i16* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP10]]
// CHECK1-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYCPY_DONE4]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1:       omp.arraycpy.done4:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], i64 [[VLA:%.*]], i16* nonnull align 2 dereferenceable(2) [[D:%.*]], i8** nonnull align 8 dereferenceable(8) [[DOTTASK_RED_:%.*]], i8** nonnull align 8 dereferenceable(8) [[DOTTASK_RED_1:%.*]]) #[[ATTR8:[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:    [[A_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[D_ADDR:%.*]] = alloca i16*, align 8
// CHECK1-NEXT:    [[DOTTASK_RED__ADDR:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    [[DOTTASK_RED__ADDR2:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 8
// CHECK1-NEXT:    [[TMP:%.*]] = alloca i32, 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* [[A]], i32** [[A_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[VLA]], i64* [[VLA_ADDR]], align 8
// CHECK1-NEXT:    store i16* [[D]], i16** [[D_ADDR]], align 8
// CHECK1-NEXT:    store i8** [[DOTTASK_RED_]], i8*** [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT:    store i8** [[DOTTASK_RED_1]], i8*** [[DOTTASK_RED__ADDR2]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i64, i64* [[VLA_ADDR]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i16*, i16** [[D_ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i8**, i8*** [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load i8**, i8*** [[DOTTASK_RED__ADDR2]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
// CHECK1-NEXT:    [[TMP7:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK1-NEXT:    [[TMP8:%.*]] = icmp ne i32 [[TMP7]], 0
// CHECK1-NEXT:    br i1 [[TMP8]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT:    store i32* [[TMP0]], i32** [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 1
// CHECK1-NEXT:    store i64 [[TMP1]], i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 2
// CHECK1-NEXT:    store i16* [[TMP2]], i16** [[TMP11]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 3
// CHECK1-NEXT:    store i8** [[TMP3]], i8*** [[TMP12]], align 8
// CHECK1-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 4
// CHECK1-NEXT:    store i8** [[TMP4]], i8*** [[TMP13]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK1-NEXT:    [[TMP14:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]], i32 1, i64 96, i64 40, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates*)* @.omp_task_entry. to i32 (i32, i8*)*))
// CHECK1-NEXT:    [[TMP15:%.*]] = bitcast i8* [[TMP14]] to %struct.kmp_task_t_with_privates*
// CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], %struct.kmp_task_t_with_privates* [[TMP15]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP18:%.*]] = load i8*, i8** [[TMP17]], align 8
// CHECK1-NEXT:    [[TMP19:%.*]] = bitcast %struct.anon* [[AGG_CAPTURED]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP18]], i8* align 8 [[TMP19]], i64 40, i1 false)
// CHECK1-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], %struct.kmp_task_t_with_privates* [[TMP15]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T:%.*]], %struct..kmp_privates.t* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP22:%.*]] = load i8*, i8** [[TMP3]], align 8
// CHECK1-NEXT:    store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T]], %struct..kmp_privates.t* [[TMP20]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP24:%.*]] = load i8*, i8** [[TMP4]], align 8
// CHECK1-NEXT:    store i8* [[TMP24]], i8** [[TMP23]], align 8
// CHECK1-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, i64* [[TMP25]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 6
// CHECK1-NEXT:    store i64 4, i64* [[TMP26]], align 8
// CHECK1-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, i64* [[TMP27]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP29:%.*]] = bitcast i8** [[TMP28]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP29]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP30:%.*]] = load i64, i64* [[TMP27]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]], i8* [[TMP14]], i32 1, i64* [[TMP25]], i64* [[TMP26]], i64 [[TMP30]], i32 1, i32 0, i64 0, i8* null)
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK1-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_privates_map.
// CHECK1-SAME: (%struct..kmp_privates.t* noalias [[TMP0:%.*]], i8*** noalias [[TMP1:%.*]], i8*** noalias [[TMP2:%.*]]) #[[ATTR9:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca %struct..kmp_privates.t*, align 8
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i8***, align 8
// CHECK1-NEXT:    [[DOTADDR2:%.*]] = alloca i8***, align 8
// CHECK1-NEXT:    store %struct..kmp_privates.t* [[TMP0]], %struct..kmp_privates.t** [[DOTADDR]], align 8
// CHECK1-NEXT:    store i8*** [[TMP1]], i8**** [[DOTADDR1]], align 8
// CHECK1-NEXT:    store i8*** [[TMP2]], i8**** [[DOTADDR2]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load %struct..kmp_privates.t*, %struct..kmp_privates.t** [[DOTADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T:%.*]], %struct..kmp_privates.t* [[TMP3]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP5:%.*]] = load i8***, i8**** [[DOTADDR1]], align 8
// CHECK1-NEXT:    store i8** [[TMP4]], i8*** [[TMP5]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T]], %struct..kmp_privates.t* [[TMP3]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP7:%.*]] = load i8***, i8**** [[DOTADDR2]], align 8
// CHECK1-NEXT:    store i8** [[TMP6]], i8*** [[TMP7]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry.
// CHECK1-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTPART_ID__ADDR_I:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTPRIVATES__ADDR_I:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTCOPY_FN__ADDR_I:%.*]] = alloca void (i8*, ...)*, align 8
// CHECK1-NEXT:    [[DOTTASK_T__ADDR_I:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[DOTLB__ADDR_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTUB__ADDR_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTST__ADDR_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTLITER__ADDR_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTREDUCTIONS__ADDR_I:%.*]] = alloca i8*, align 8
// CHECK1-NEXT:    [[__CONTEXT_ADDR_I:%.*]] = alloca %struct.anon*, align 8
// CHECK1-NEXT:    [[DOTFIRSTPRIV_PTR_ADDR_I:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    [[DOTFIRSTPRIV_PTR_ADDR1_I:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca %struct.kmp_task_t_with_privates*, align 8
// CHECK1-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    store %struct.kmp_task_t_with_privates* [[TMP1]], %struct.kmp_task_t_with_privates** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates*, %struct.kmp_task_t_with_privates** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], %struct.kmp_task_t_with_privates* [[TMP3]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 2
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to %struct.anon*
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], %struct.kmp_task_t_with_privates* [[TMP3]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP10:%.*]] = bitcast %struct..kmp_privates.t* [[TMP9]] to i8*
// CHECK1-NEXT:    [[TMP11:%.*]] = bitcast %struct.kmp_task_t_with_privates* [[TMP3]] to i8*
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP13:%.*]] = load i64, i64* [[TMP12]], align 8
// CHECK1-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP15:%.*]] = load i64, i64* [[TMP14]], align 8
// CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 7
// CHECK1-NEXT:    [[TMP17:%.*]] = load i64, i64* [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 8
// CHECK1-NEXT:    [[TMP19:%.*]] = load i32, i32* [[TMP18]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP21:%.*]] = load i8*, i8** [[TMP20]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META2:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !13
// CHECK1-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store i8* [[TMP10]], i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store void (i8*, ...)* bitcast (void (%struct..kmp_privates.t*, i8***, i8***)* @.omp_task_privates_map. to void (i8*, ...)*), void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store i8* [[TMP11]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store i64 [[TMP13]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store i64 [[TMP15]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store i64 [[TMP17]], i64* [[DOTST__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store i32 [[TMP19]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !13
// CHECK1-NEXT:    store i8* [[TMP21]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    store %struct.anon* [[TMP8]], %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[TMP22:%.*]] = load %struct.anon*, %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_ANON:%.*]], %struct.anon* [[TMP22]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP24:%.*]] = load i64, i64* [[TMP23]], align 8
// CHECK1-NEXT:    [[TMP25:%.*]] = load void (i8*, ...)*, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[TMP26:%.*]] = load i8*, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[TMP27:%.*]] = bitcast void (i8*, ...)* [[TMP25]] to void (i8*, i8***, i8***)*
// CHECK1-NEXT:    call void [[TMP27]](i8* [[TMP26]], i8*** [[DOTFIRSTPRIV_PTR_ADDR_I]], i8*** [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
// CHECK1-NEXT:    [[TMP28:%.*]] = load i8**, i8*** [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[TMP29:%.*]] = load i8**, i8*** [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[TMP22]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP31:%.*]] = load i32*, i32** [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP32:%.*]] = load i8*, i8** [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP33:%.*]] = load i32, i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !13
// CHECK1-NEXT:    [[TMP34:%.*]] = bitcast i32* [[TMP31]] to i8*
// CHECK1-NEXT:    [[TMP35:%.*]] = call i8* @__kmpc_task_reduction_get_th_data(i32 [[TMP33]], i8* [[TMP32]], i8* [[TMP34]]) #[[ATTR3]]
// CHECK1-NEXT:    [[CONV_I:%.*]] = bitcast i8* [[TMP35]] to i32*
// CHECK1-NEXT:    [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[TMP22]], i32 0, i32 2
// CHECK1-NEXT:    [[TMP37:%.*]] = load i16*, i16** [[TMP36]], align 8
// CHECK1-NEXT:    [[TMP38:%.*]] = mul nuw i64 [[TMP24]], 2
// CHECK1-NEXT:    [[TMP39:%.*]] = udiv exact i64 [[TMP38]], ptrtoint (i16* getelementptr (i16, i16* null, i32 1) to i64)
// CHECK1-NEXT:    [[TMP40:%.*]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @[[GLOB1]], i32 [[TMP33]], i8* bitcast (i64* @{{reduction_size[.].+[.]}}) #[[ATTR3]]
// CHECK1-NEXT:    [[TMP41:%.*]] = bitcast i8* [[TMP40]] to i64*
// CHECK1-NEXT:    store i64 [[TMP39]], i64* [[TMP41]], align 8
// CHECK1-NEXT:    [[TMP42:%.*]] = load i8*, i8** [[TMP29]], align 8
// CHECK1-NEXT:    [[TMP43:%.*]] = bitcast i16* [[TMP37]] to i8*
// CHECK1-NEXT:    [[TMP44:%.*]] = call i8* @__kmpc_task_reduction_get_th_data(i32 [[TMP33]], i8* [[TMP42]], i8* [[TMP43]]) #[[ATTR3]]
// CHECK1-NEXT:    [[CONV2_I:%.*]] = bitcast i8* [[TMP44]] to i16*
// CHECK1-NEXT:    [[TMP45:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[CONV3_I:%.*]] = trunc i64 [[TMP45]] to i32
// CHECK1-NEXT:    store i32 [[CONV3_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP46:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK1-NEXT:    [[CONV4_I:%.*]] = sext i32 [[TMP46]] to i64
// CHECK1-NEXT:    [[TMP47:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !13
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV4_I]], [[TMP47]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__9_EXIT:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP48:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK1-NEXT:    store i32 [[TMP48]], i32* [[I_I]], align 4, !noalias !13
// CHECK1-NEXT:    [[TMP49:%.*]] = load i32, i32* [[CONV_I]], align 4
// CHECK1-NEXT:    [[IDXPROM_I:%.*]] = sext i32 [[TMP49]] to i64
// CHECK1-NEXT:    [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, i16* [[CONV2_I]], i64 [[IDXPROM_I]]
// CHECK1-NEXT:    [[TMP50:%.*]] = load i16, i16* [[ARRAYIDX_I]], align 2
// CHECK1-NEXT:    [[CONV5_I:%.*]] = sext i16 [[TMP50]] to i32
// CHECK1-NEXT:    [[TMP51:%.*]] = load i32, i32* [[CONV_I]], align 4
// CHECK1-NEXT:    [[ADD6_I:%.*]] = add nsw i32 [[TMP51]], [[CONV5_I]]
// CHECK1-NEXT:    store i32 [[ADD6_I]], i32* [[CONV_I]], align 4
// CHECK1-NEXT:    [[TMP52:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK1-NEXT:    [[ADD7_I:%.*]] = add nsw i32 [[TMP52]], 1
// CHECK1-NEXT:    store i32 [[ADD7_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK1:       .omp_outlined..9.exit:
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC2Ev
// CHECK1-SAME: (%struct.S* nonnull 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:    [[A:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK1-NEXT:    store i32 0, i32* [[A]], align 4
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SD2Ev
// CHECK1-SAME: (%struct.S* nonnull 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 {{[^@]+}}@_ZN1SC1ERKS_
// CHECK1-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) unnamed_addr #[[ATTR1]] 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:    [[TMP1:%.*]] = load %struct.S*, %struct.S** [[DOTADDR]], align 8
// CHECK1-NEXT:    call void @_ZN1SC2ERKS_(%struct.S* nonnull dereferenceable(4) [[THIS1]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP1]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC2ERKS_
// CHECK1-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) unnamed_addr #[[ATTR1]] 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 void
//
//
// CHECK2-LABEL: define {{[^@]+}}@main
// CHECK2-SAME: (i32 [[ARGC:%.*]], i8** [[ARGV:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[ARGV_ADDR:%.*]] = alloca i8**, align 8
// CHECK2-NEXT:    [[A:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[B:%.*]] = alloca float, align 4
// CHECK2-NEXT:    [[C:%.*]] = alloca [5 x %struct.S], align 16
// CHECK2-NEXT:    [[SAVED_STACK:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTRD_INPUT_:%.*]] = alloca [3 x %struct.kmp_taskred_input_t], align 8
// CHECK2-NEXT:    [[DOTTASK_RED_:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTRD_INPUT_3:%.*]] = alloca [2 x %struct.kmp_taskred_input_t.0], align 8
// CHECK2-NEXT:    [[DOTTASK_RED_6:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1:[0-9]+]])
// CHECK2-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    store i32 [[ARGC]], i32* [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i8** [[ARGV]], i8*** [[ARGV_ADDR]], align 8
// CHECK2-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK2-NEXT:    [[ARRAYCTOR_END:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK2-NEXT:    br label [[ARRAYCTOR_LOOP:%.*]]
// CHECK2:       arrayctor.loop:
// CHECK2-NEXT:    [[ARRAYCTOR_CUR:%.*]] = phi %struct.S* [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[ARRAYCTOR_NEXT:%.*]], [[ARRAYCTOR_LOOP]] ]
// CHECK2-NEXT:    call void @_ZN1SC1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYCTOR_CUR]])
// CHECK2-NEXT:    [[ARRAYCTOR_NEXT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYCTOR_CUR]], i64 1
// CHECK2-NEXT:    [[ARRAYCTOR_DONE:%.*]] = icmp eq %struct.S* [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK2-NEXT:    br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK2:       arrayctor.cont:
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = zext i32 [[TMP1]] to i64
// CHECK2-NEXT:    [[TMP3:%.*]] = call i8* @llvm.stacksave()
// CHECK2-NEXT:    store i8* [[TMP3]], i8** [[SAVED_STACK]], align 8
// CHECK2-NEXT:    [[VLA:%.*]] = alloca i16, i64 [[TMP2]], align 16
// CHECK2-NEXT:    store i64 [[TMP2]], i64* [[__VLA_EXPR0]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK2-NEXT:    [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T:%.*]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i32* [[A]] to i8*
// CHECK2-NEXT:    store i8* [[TMP5]], i8** [[TMP4]], align 8
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP7:%.*]] = bitcast i32* [[A]] to i8*
// CHECK2-NEXT:    store i8* [[TMP7]], i8** [[TMP6]], align 8
// CHECK2-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 2
// CHECK2-NEXT:    store i64 4, i64* [[TMP8]], align 8
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 3
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init. to i8*), i8** [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 4
// CHECK2-NEXT:    store i8* null, i8** [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 5
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb. to i8*), i8** [[TMP11]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP13:%.*]] = bitcast i32* [[TMP12]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP13]], i8 0, i64 4, i1 false)
// CHECK2-NEXT:    [[DOTRD_INPUT_GEP_1:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP15:%.*]] = bitcast float* [[B]] to i8*
// CHECK2-NEXT:    store i8* [[TMP15]], i8** [[TMP14]], align 8
// CHECK2-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP17:%.*]] = bitcast float* [[B]] to i8*
// CHECK2-NEXT:    store i8* [[TMP17]], i8** [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 2
// CHECK2-NEXT:    store i64 4, i64* [[TMP18]], align 8
// CHECK2-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 3
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..1 to i8*), i8** [[TMP19]], align 8
// CHECK2-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 4
// CHECK2-NEXT:    store i8* null, i8** [[TMP20]], align 8
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 5
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..2 to i8*), i8** [[TMP21]], align 8
// CHECK2-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_1]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP23:%.*]] = bitcast i32* [[TMP22]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP23]], i8 0, i64 4, i1 false)
// CHECK2-NEXT:    [[DOTRD_INPUT_GEP_2:%.*]] = getelementptr inbounds [3 x %struct.kmp_taskred_input_t], [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]], i64 0, i64 2
// CHECK2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP25:%.*]] = bitcast i32* [[ARGC_ADDR]] to i8*
// CHECK2-NEXT:    store i8* [[TMP25]], i8** [[TMP24]], align 8
// CHECK2-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP27:%.*]] = bitcast i32* [[ARGC_ADDR]] to i8*
// CHECK2-NEXT:    store i8* [[TMP27]], i8** [[TMP26]], align 8
// CHECK2-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 2
// CHECK2-NEXT:    store i64 4, i64* [[TMP28]], align 8
// CHECK2-NEXT:    [[TMP29:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 3
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..3 to i8*), i8** [[TMP29]], align 8
// CHECK2-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 4
// CHECK2-NEXT:    store i8* null, i8** [[TMP30]], align 8
// CHECK2-NEXT:    [[TMP31:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 5
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..4 to i8*), i8** [[TMP31]], align 8
// CHECK2-NEXT:    [[TMP32:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T]], %struct.kmp_taskred_input_t* [[DOTRD_INPUT_GEP_2]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP33:%.*]] = bitcast i32* [[TMP32]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP33]], i8 0, i64 4, i1 false)
// CHECK2-NEXT:    [[TMP34:%.*]] = bitcast [3 x %struct.kmp_taskred_input_t]* [[DOTRD_INPUT_]] to i8*
// CHECK2-NEXT:    [[TMP35:%.*]] = call i8* @__kmpc_taskred_init(i32 [[TMP0]], i32 3, i8* [[TMP34]])
// CHECK2-NEXT:    store i8* [[TMP35]], i8** [[DOTTASK_RED_]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK2-NEXT:    [[DOTRD_INPUT_GEP_4:%.*]] = getelementptr inbounds [2 x %struct.kmp_taskred_input_t.0], [2 x %struct.kmp_taskred_input_t.0]* [[DOTRD_INPUT_3]], i64 0, i64 0
// CHECK2-NEXT:    [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0:%.*]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP37:%.*]] = bitcast [5 x %struct.S]* [[C]] to i8*
// CHECK2-NEXT:    store i8* [[TMP37]], i8** [[TMP36]], align 8
// CHECK2-NEXT:    [[TMP38:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP39:%.*]] = bitcast [5 x %struct.S]* [[C]] to i8*
// CHECK2-NEXT:    store i8* [[TMP39]], i8** [[TMP38]], align 8
// CHECK2-NEXT:    [[TMP40:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK2-NEXT:    store i64 20, i64* [[TMP40]], align 8
// CHECK2-NEXT:    [[TMP41:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..5 to i8*), i8** [[TMP41]], align 8
// CHECK2-NEXT:    [[TMP42:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 4
// CHECK2-NEXT:    store i8* bitcast (void (i8*)* @.red_fini. to i8*), i8** [[TMP42]], align 8
// CHECK2-NEXT:    [[TMP43:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 5
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..6 to i8*), i8** [[TMP43]], align 8
// CHECK2-NEXT:    [[TMP44:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_4]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP45:%.*]] = bitcast i32* [[TMP44]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP45]], i8 0, i64 4, i1 false)
// CHECK2-NEXT:    [[DOTRD_INPUT_GEP_5:%.*]] = getelementptr inbounds [2 x %struct.kmp_taskred_input_t.0], [2 x %struct.kmp_taskred_input_t.0]* [[DOTRD_INPUT_3]], i64 0, i64 1
// CHECK2-NEXT:    [[TMP46:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP47:%.*]] = bitcast i16* [[VLA]] to i8*
// CHECK2-NEXT:    store i8* [[TMP47]], i8** [[TMP46]], align 8
// CHECK2-NEXT:    [[TMP48:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP49:%.*]] = bitcast i16* [[VLA]] to i8*
// CHECK2-NEXT:    store i8* [[TMP49]], i8** [[TMP48]], align 8
// CHECK2-NEXT:    [[TMP50:%.*]] = mul nuw i64 [[TMP2]], 2
// CHECK2-NEXT:    [[TMP51:%.*]] = udiv exact i64 [[TMP50]], ptrtoint (i16* getelementptr (i16, i16* null, i32 1) to i64)
// CHECK2-NEXT:    [[TMP52:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK2-NEXT:    store i64 [[TMP50]], i64* [[TMP52]], align 8
// CHECK2-NEXT:    [[TMP53:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_init..7 to i8*), i8** [[TMP53]], align 8
// CHECK2-NEXT:    [[TMP54:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 4
// CHECK2-NEXT:    store i8* null, i8** [[TMP54]], align 8
// CHECK2-NEXT:    [[TMP55:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 5
// CHECK2-NEXT:    store i8* bitcast (void (i8*, i8*)* @.red_comb..8 to i8*), i8** [[TMP55]], align 8
// CHECK2-NEXT:    [[TMP56:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASKRED_INPUT_T_0]], %struct.kmp_taskred_input_t.0* [[DOTRD_INPUT_GEP_5]], i32 0, i32 6
// CHECK2-NEXT:    store i32 1, i32* [[TMP56]], align 8
// CHECK2-NEXT:    [[TMP57:%.*]] = bitcast [2 x %struct.kmp_taskred_input_t.0]* [[DOTRD_INPUT_3]] to i8*
// CHECK2-NEXT:    [[TMP58:%.*]] = call i8* @__kmpc_taskred_init(i32 [[TMP0]], i32 2, i8* [[TMP57]])
// CHECK2-NEXT:    store i8* [[TMP58]], i8** [[DOTTASK_RED_6]], align 8
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i16*, i8**, i8**)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* [[A]], i64 [[TMP2]], i16* [[VLA]], i8** [[DOTTASK_RED_]], i8** [[DOTTASK_RED_6]])
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK2-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    [[TMP59:%.*]] = load i8*, i8** [[SAVED_STACK]], align 8
// CHECK2-NEXT:    call void @llvm.stackrestore(i8* [[TMP59]])
// CHECK2-NEXT:    [[ARRAY_BEGIN7:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP60:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAY_BEGIN7]], i64 5
// CHECK2-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK2:       arraydestroy.body:
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP60]], [[ARRAYCTOR_CONT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK2-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3:[0-9]+]]
// CHECK2-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN7]]
// CHECK2-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE8:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK2:       arraydestroy.done8:
// CHECK2-NEXT:    [[TMP61:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    ret i32 [[TMP61]]
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SC1Ev
// CHECK2-SAME: (%struct.S* nonnull 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 @_ZN1SC2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_init.
// CHECK2-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 i32*
// CHECK2-NEXT:    store i32 0, i32* [[TMP3]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_comb.
// 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:    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 i32*
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to i32*
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP3]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP5]], align 8
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP6]], [[TMP7]]
// CHECK2-NEXT:    store i32 [[ADD]], i32* [[TMP3]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_init..1
// CHECK2-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 float*
// CHECK2-NEXT:    store float 0.000000e+00, float* [[TMP3]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_comb..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:    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 float*
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to float*
// CHECK2-NEXT:    [[TMP6:%.*]] = load float, float* [[TMP3]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = load float, float* [[TMP5]], align 8
// CHECK2-NEXT:    [[ADD:%.*]] = fadd float [[TMP6]], [[TMP7]]
// CHECK2-NEXT:    store float [[ADD]], float* [[TMP3]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_init..3
// CHECK2-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 i32*
// CHECK2-NEXT:    store i32 0, i32* [[TMP3]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_comb..4
// 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:    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 i32*
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to i32*
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP3]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP5]], align 8
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP6]], [[TMP7]]
// CHECK2-NEXT:    store i32 [[ADD]], i32* [[TMP3]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_init..5
// CHECK2-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// 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 [5 x %struct.S]*
// CHECK2-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[TMP3]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK2-NEXT:    [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq %struct.S* [[ARRAY_BEGIN]], [[TMP4]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK2:       omp.arrayinit.body:
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi %struct.S* [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK2-NEXT:    call void @_ZN1SC1Ev(%struct.S* nonnull dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]])
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], %struct.S* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq %struct.S* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP4]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK2:       omp.arrayinit.done:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_fini.
// CHECK2-SAME: (i8* [[TMP0:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = bitcast i8* [[TMP1]] to [5 x %struct.S]*
// CHECK2-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[TMP2]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK2-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK2:       arraydestroy.body:
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP3]], [[ENTRY:%.*]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK2-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK2-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]]
// 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:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SD1Ev
// CHECK2-SAME: (%struct.S* nonnull 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 @_ZN1SD2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]]) #[[ATTR3]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_comb..6
// 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:%.*]], 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 %struct.S*
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to %struct.S*
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr [[STRUCT_S]], %struct.S* [[TMP3]], i64 5
// CHECK2-NEXT:    [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq %struct.S* [[TMP3]], [[TMP6]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE2:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK2:       omp.arraycpy.body:
// CHECK2-NEXT:    [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP5]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP3]], [[ENTRY]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK2-NEXT:    call void @_ZplRK1SS1_(%struct.S* sret([[STRUCT_S]]) align 4 [[REF_TMP]], %struct.S* nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]], %struct.S* nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_SRCELEMENTPAST]])
// CHECK2-NEXT:    [[CALL:%.*]] = call nonnull align 4 dereferenceable(4) %struct.S* @_ZN1SaSERKS_(%struct.S* nonnull dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]], %struct.S* nonnull align 4 dereferenceable(4) [[REF_TMP]])
// CHECK2-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[REF_TMP]]) #[[ATTR3]]
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], %struct.S* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK2-NEXT:    [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr [[STRUCT_S]], %struct.S* [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq %struct.S* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP6]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYCPY_DONE2]], label [[OMP_ARRAYCPY_BODY]]
// CHECK2:       omp.arraycpy.done2:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZplRK1SS1_
// CHECK2-SAME: (%struct.S* noalias sret([[STRUCT_S:%.*]]) align 4 [[AGG_RESULT:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[A:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[B:%.*]]) #[[ATTR7:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[RESULT_PTR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[B_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = bitcast %struct.S* [[AGG_RESULT]] to i8*
// CHECK2-NEXT:    store i8* [[TMP0]], i8** [[RESULT_PTR]], align 8
// CHECK2-NEXT:    store %struct.S* [[A]], %struct.S** [[A_ADDR]], align 8
// CHECK2-NEXT:    store %struct.S* [[B]], %struct.S** [[B_ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load %struct.S*, %struct.S** [[A_ADDR]], align 8
// CHECK2-NEXT:    call void @_ZN1SC1ERKS_(%struct.S* nonnull dereferenceable(4) [[AGG_RESULT]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP1]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SaSERKS_
// CHECK2-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) #[[ATTR7]] 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 {{[^@]+}}@.red_init..7
// CHECK2-SAME: (i8* noalias [[TMP0:%.*]], i8* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i8* bitcast (i64* @{{reduction_size[.].+[.]}})
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to i64*
// CHECK2-NEXT:    [[TMP6:%.*]] = load i64, i64* [[TMP5]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = bitcast i8* [[TMP3]] to i16*
// CHECK2-NEXT:    [[TMP8:%.*]] = getelementptr i16, i16* [[TMP7]], i64 [[TMP6]]
// CHECK2-NEXT:    [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq i16* [[TMP7]], [[TMP8]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK2:       omp.arrayinit.body:
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi i16* [ [[TMP7]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK2-NEXT:    store i16 0, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq i16* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP8]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK2:       omp.arrayinit.done:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.red_comb..8
// 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:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT:    store i8* [[TMP0]], i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    store i8* [[TMP1]], i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i8* bitcast (i64* @{{reduction_size[.].+[.]}})
// CHECK2-NEXT:    [[TMP4:%.*]] = bitcast i8* [[TMP3]] to i64*
// CHECK2-NEXT:    [[TMP5:%.*]] = load i64, i64* [[TMP4]], align 8
// CHECK2-NEXT:    [[TMP6:%.*]] = load i8*, i8** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = bitcast i8* [[TMP6]] to i16*
// CHECK2-NEXT:    [[TMP8:%.*]] = load i8*, i8** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP9:%.*]] = bitcast i8* [[TMP8]] to i16*
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr i16, i16* [[TMP7]], i64 [[TMP5]]
// CHECK2-NEXT:    [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq i16* [[TMP7]], [[TMP10]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE4:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK2:       omp.arraycpy.body:
// CHECK2-NEXT:    [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi i16* [ [[TMP9]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi i16* [ [[TMP7]], [[ENTRY]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK2-NEXT:    [[TMP11:%.*]] = load i16, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK2-NEXT:    [[CONV:%.*]] = sext i16 [[TMP11]] to i32
// CHECK2-NEXT:    [[TMP12:%.*]] = load i16, i16* [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
// CHECK2-NEXT:    [[CONV2:%.*]] = sext i16 [[TMP12]] to i32
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], [[CONV2]]
// CHECK2-NEXT:    [[CONV3:%.*]] = trunc i32 [[ADD]] to i16
// CHECK2-NEXT:    store i16 [[CONV3]], i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, i16* [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
// CHECK2-NEXT:    [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i16, i16* [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK2-NEXT:    [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq i16* [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP10]]
// CHECK2-NEXT:    br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYCPY_DONE4]], label [[OMP_ARRAYCPY_BODY]]
// CHECK2:       omp.arraycpy.done4:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], i64 [[VLA:%.*]], i16* nonnull align 2 dereferenceable(2) [[D:%.*]], i8** nonnull align 8 dereferenceable(8) [[DOTTASK_RED_:%.*]], i8** nonnull align 8 dereferenceable(8) [[DOTTASK_RED_1:%.*]]) #[[ATTR8:[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:    [[A_ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[D_ADDR:%.*]] = alloca i16*, align 8
// CHECK2-NEXT:    [[DOTTASK_RED__ADDR:%.*]] = alloca i8**, align 8
// CHECK2-NEXT:    [[DOTTASK_RED__ADDR2:%.*]] = alloca i8**, align 8
// CHECK2-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 8
// CHECK2-NEXT:    [[TMP:%.*]] = alloca i32, 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* [[A]], i32** [[A_ADDR]], align 8
// CHECK2-NEXT:    store i64 [[VLA]], i64* [[VLA_ADDR]], align 8
// CHECK2-NEXT:    store i16* [[D]], i16** [[D_ADDR]], align 8
// CHECK2-NEXT:    store i8** [[DOTTASK_RED_]], i8*** [[DOTTASK_RED__ADDR]], align 8
// CHECK2-NEXT:    store i8** [[DOTTASK_RED_1]], i8*** [[DOTTASK_RED__ADDR2]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i64, i64* [[VLA_ADDR]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i16*, i16** [[D_ADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load i8**, i8*** [[DOTTASK_RED__ADDR]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = load i8**, i8*** [[DOTTASK_RED__ADDR2]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
// CHECK2-NEXT:    [[TMP7:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK2-NEXT:    [[TMP8:%.*]] = icmp ne i32 [[TMP7]], 0
// CHECK2-NEXT:    br i1 [[TMP8]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 0
// CHECK2-NEXT:    store i32* [[TMP0]], i32** [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 1
// CHECK2-NEXT:    store i64 [[TMP1]], i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 2
// CHECK2-NEXT:    store i16* [[TMP2]], i16** [[TMP11]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 3
// CHECK2-NEXT:    store i8** [[TMP3]], i8*** [[TMP12]], align 8
// CHECK2-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[AGG_CAPTURED]], i32 0, i32 4
// CHECK2-NEXT:    store i8** [[TMP4]], i8*** [[TMP13]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK2-NEXT:    [[TMP14:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]], i32 1, i64 96, i64 40, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates*)* @.omp_task_entry. to i32 (i32, i8*)*))
// CHECK2-NEXT:    [[TMP15:%.*]] = bitcast i8* [[TMP14]] to %struct.kmp_task_t_with_privates*
// CHECK2-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], %struct.kmp_task_t_with_privates* [[TMP15]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP18:%.*]] = load i8*, i8** [[TMP17]], align 8
// CHECK2-NEXT:    [[TMP19:%.*]] = bitcast %struct.anon* [[AGG_CAPTURED]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP18]], i8* align 8 [[TMP19]], i64 40, i1 false)
// CHECK2-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], %struct.kmp_task_t_with_privates* [[TMP15]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T:%.*]], %struct..kmp_privates.t* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP22:%.*]] = load i8*, i8** [[TMP3]], align 8
// CHECK2-NEXT:    store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK2-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T]], %struct..kmp_privates.t* [[TMP20]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP24:%.*]] = load i8*, i8** [[TMP4]], align 8
// CHECK2-NEXT:    store i8* [[TMP24]], i8** [[TMP23]], align 8
// CHECK2-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 5
// CHECK2-NEXT:    store i64 0, i64* [[TMP25]], align 8
// CHECK2-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 6
// CHECK2-NEXT:    store i64 4, i64* [[TMP26]], align 8
// CHECK2-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 7
// CHECK2-NEXT:    store i64 1, i64* [[TMP27]], align 8
// CHECK2-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP16]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP29:%.*]] = bitcast i8** [[TMP28]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP29]], i8 0, i64 8, i1 false)
// CHECK2-NEXT:    [[TMP30:%.*]] = load i64, i64* [[TMP27]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]], i8* [[TMP14]], i32 1, i64* [[TMP25]], i64* [[TMP26]], i64 [[TMP30]], i32 1, i32 0, i64 0, i8* null)
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK2-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP6]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_privates_map.
// CHECK2-SAME: (%struct..kmp_privates.t* noalias [[TMP0:%.*]], i8*** noalias [[TMP1:%.*]], i8*** noalias [[TMP2:%.*]]) #[[ATTR9:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca %struct..kmp_privates.t*, align 8
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i8***, align 8
// CHECK2-NEXT:    [[DOTADDR2:%.*]] = alloca i8***, align 8
// CHECK2-NEXT:    store %struct..kmp_privates.t* [[TMP0]], %struct..kmp_privates.t** [[DOTADDR]], align 8
// CHECK2-NEXT:    store i8*** [[TMP1]], i8**** [[DOTADDR1]], align 8
// CHECK2-NEXT:    store i8*** [[TMP2]], i8**** [[DOTADDR2]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load %struct..kmp_privates.t*, %struct..kmp_privates.t** [[DOTADDR]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T:%.*]], %struct..kmp_privates.t* [[TMP3]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP5:%.*]] = load i8***, i8**** [[DOTADDR1]], align 8
// CHECK2-NEXT:    store i8** [[TMP4]], i8*** [[TMP5]], align 8
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT__KMP_PRIVATES_T]], %struct..kmp_privates.t* [[TMP3]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP7:%.*]] = load i8***, i8**** [[DOTADDR2]], align 8
// CHECK2-NEXT:    store i8** [[TMP6]], i8*** [[TMP7]], align 8
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_entry.
// CHECK2-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates* noalias [[TMP1:%.*]]) #[[ATTR5]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTPART_ID__ADDR_I:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTPRIVATES__ADDR_I:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTCOPY_FN__ADDR_I:%.*]] = alloca void (i8*, ...)*, align 8
// CHECK2-NEXT:    [[DOTTASK_T__ADDR_I:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[DOTLB__ADDR_I:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTUB__ADDR_I:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTST__ADDR_I:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTLITER__ADDR_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTREDUCTIONS__ADDR_I:%.*]] = alloca i8*, align 8
// CHECK2-NEXT:    [[__CONTEXT_ADDR_I:%.*]] = alloca %struct.anon*, align 8
// CHECK2-NEXT:    [[DOTFIRSTPRIV_PTR_ADDR_I:%.*]] = alloca i8**, align 8
// CHECK2-NEXT:    [[DOTFIRSTPRIV_PTR_ADDR1_I:%.*]] = alloca i8**, align 8
// CHECK2-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca %struct.kmp_task_t_with_privates*, align 8
// CHECK2-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    store %struct.kmp_task_t_with_privates* [[TMP1]], %struct.kmp_task_t_with_privates** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates*, %struct.kmp_task_t_with_privates** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], %struct.kmp_task_t_with_privates* [[TMP3]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 2
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP7:%.*]] = load i8*, i8** [[TMP6]], align 8
// CHECK2-NEXT:    [[TMP8:%.*]] = bitcast i8* [[TMP7]] to %struct.anon*
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], %struct.kmp_task_t_with_privates* [[TMP3]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP10:%.*]] = bitcast %struct..kmp_privates.t* [[TMP9]] to i8*
// CHECK2-NEXT:    [[TMP11:%.*]] = bitcast %struct.kmp_task_t_with_privates* [[TMP3]] to i8*
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK2-NEXT:    [[TMP13:%.*]] = load i64, i64* [[TMP12]], align 8
// CHECK2-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP15:%.*]] = load i64, i64* [[TMP14]], align 8
// CHECK2-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 7
// CHECK2-NEXT:    [[TMP17:%.*]] = load i64, i64* [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 8
// CHECK2-NEXT:    [[TMP19:%.*]] = load i32, i32* [[TMP18]], align 8
// CHECK2-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP21:%.*]] = load i8*, i8** [[TMP20]], align 8
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META2:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !13
// CHECK2-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store i8* [[TMP10]], i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store void (i8*, ...)* bitcast (void (%struct..kmp_privates.t*, i8***, i8***)* @.omp_task_privates_map. to void (i8*, ...)*), void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store i8* [[TMP11]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store i64 [[TMP13]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store i64 [[TMP15]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store i64 [[TMP17]], i64* [[DOTST__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store i32 [[TMP19]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !13
// CHECK2-NEXT:    store i8* [[TMP21]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    store %struct.anon* [[TMP8]], %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[TMP22:%.*]] = load %struct.anon*, %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_ANON:%.*]], %struct.anon* [[TMP22]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP24:%.*]] = load i64, i64* [[TMP23]], align 8
// CHECK2-NEXT:    [[TMP25:%.*]] = load void (i8*, ...)*, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[TMP26:%.*]] = load i8*, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[TMP27:%.*]] = bitcast void (i8*, ...)* [[TMP25]] to void (i8*, i8***, i8***)*
// CHECK2-NEXT:    call void [[TMP27]](i8* [[TMP26]], i8*** [[DOTFIRSTPRIV_PTR_ADDR_I]], i8*** [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
// CHECK2-NEXT:    [[TMP28:%.*]] = load i8**, i8*** [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[TMP29:%.*]] = load i8**, i8*** [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[TMP22]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP31:%.*]] = load i32*, i32** [[TMP30]], align 8
// CHECK2-NEXT:    [[TMP32:%.*]] = load i8*, i8** [[TMP28]], align 8
// CHECK2-NEXT:    [[TMP33:%.*]] = load i32, i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !13
// CHECK2-NEXT:    [[TMP34:%.*]] = bitcast i32* [[TMP31]] to i8*
// CHECK2-NEXT:    [[TMP35:%.*]] = call i8* @__kmpc_task_reduction_get_th_data(i32 [[TMP33]], i8* [[TMP32]], i8* [[TMP34]]) #[[ATTR3]]
// CHECK2-NEXT:    [[CONV_I:%.*]] = bitcast i8* [[TMP35]] to i32*
// CHECK2-NEXT:    [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], %struct.anon* [[TMP22]], i32 0, i32 2
// CHECK2-NEXT:    [[TMP37:%.*]] = load i16*, i16** [[TMP36]], align 8
// CHECK2-NEXT:    [[TMP38:%.*]] = mul nuw i64 [[TMP24]], 2
// CHECK2-NEXT:    [[TMP39:%.*]] = udiv exact i64 [[TMP38]], ptrtoint (i16* getelementptr (i16, i16* null, i32 1) to i64)
// CHECK2-NEXT:    [[TMP40:%.*]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @[[GLOB1]], i32 [[TMP33]], i8* bitcast (i64* @{{reduction_size[.].+[.]}}) #[[ATTR3]]
// CHECK2-NEXT:    [[TMP41:%.*]] = bitcast i8* [[TMP40]] to i64*
// CHECK2-NEXT:    store i64 [[TMP39]], i64* [[TMP41]], align 8
// CHECK2-NEXT:    [[TMP42:%.*]] = load i8*, i8** [[TMP29]], align 8
// CHECK2-NEXT:    [[TMP43:%.*]] = bitcast i16* [[TMP37]] to i8*
// CHECK2-NEXT:    [[TMP44:%.*]] = call i8* @__kmpc_task_reduction_get_th_data(i32 [[TMP33]], i8* [[TMP42]], i8* [[TMP43]]) #[[ATTR3]]
// CHECK2-NEXT:    [[CONV2_I:%.*]] = bitcast i8* [[TMP44]] to i16*
// CHECK2-NEXT:    [[TMP45:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[CONV3_I:%.*]] = trunc i64 [[TMP45]] to i32
// CHECK2-NEXT:    store i32 [[CONV3_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK2:       omp.inner.for.cond.i:
// CHECK2-NEXT:    [[TMP46:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK2-NEXT:    [[CONV4_I:%.*]] = sext i32 [[TMP46]] to i64
// CHECK2-NEXT:    [[TMP47:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !13
// CHECK2-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV4_I]], [[TMP47]]
// CHECK2-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__9_EXIT:%.*]]
// CHECK2:       omp.inner.for.body.i:
// CHECK2-NEXT:    [[TMP48:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK2-NEXT:    store i32 [[TMP48]], i32* [[I_I]], align 4, !noalias !13
// CHECK2-NEXT:    [[TMP49:%.*]] = load i32, i32* [[CONV_I]], align 4
// CHECK2-NEXT:    [[IDXPROM_I:%.*]] = sext i32 [[TMP49]] to i64
// CHECK2-NEXT:    [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, i16* [[CONV2_I]], i64 [[IDXPROM_I]]
// CHECK2-NEXT:    [[TMP50:%.*]] = load i16, i16* [[ARRAYIDX_I]], align 2
// CHECK2-NEXT:    [[CONV5_I:%.*]] = sext i16 [[TMP50]] to i32
// CHECK2-NEXT:    [[TMP51:%.*]] = load i32, i32* [[CONV_I]], align 4
// CHECK2-NEXT:    [[ADD6_I:%.*]] = add nsw i32 [[TMP51]], [[CONV5_I]]
// CHECK2-NEXT:    store i32 [[ADD6_I]], i32* [[CONV_I]], align 4
// CHECK2-NEXT:    [[TMP52:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK2-NEXT:    [[ADD7_I:%.*]] = add nsw i32 [[TMP52]], 1
// CHECK2-NEXT:    store i32 [[ADD7_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !13
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK2:       .omp_outlined..9.exit:
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SC2Ev
// CHECK2-SAME: (%struct.S* nonnull 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:    [[A:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK2-NEXT:    store i32 0, i32* [[A]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SD2Ev
// CHECK2-SAME: (%struct.S* nonnull 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 {{[^@]+}}@_ZN1SC1ERKS_
// CHECK2-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) unnamed_addr #[[ATTR1]] 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:    [[TMP1:%.*]] = load %struct.S*, %struct.S** [[DOTADDR]], align 8
// CHECK2-NEXT:    call void @_ZN1SC2ERKS_(%struct.S* nonnull dereferenceable(4) [[THIS1]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP1]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SC2ERKS_
// CHECK2-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]], %struct.S* nonnull align 4 dereferenceable(4) [[TMP0:%.*]]) unnamed_addr #[[ATTR1]] 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 void
//
//
// CHECK3-LABEL: define {{[^@]+}}@main
// CHECK3-SAME: (i32 [[ARGC:%.*]], i8** [[ARGV:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[ARGV_ADDR:%.*]] = alloca i8**, align 8
// CHECK3-NEXT:    [[A:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[B:%.*]] = alloca float, align 4
// CHECK3-NEXT:    [[C:%.*]] = alloca [5 x %struct.S], align 16
// CHECK3-NEXT:    [[SAVED_STACK:%.*]] = alloca i8*, align 8
// CHECK3-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8
// CHECK3-NEXT:    [[DOTTASK_RED_:%.*]] = alloca i8*, align 8
// CHECK3-NEXT:    [[DOTTASK_RED_1:%.*]] = alloca i8*, align 8
// CHECK3-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK3-NEXT:    store i32 [[ARGC]], i32* [[ARGC_ADDR]], align 4
// CHECK3-NEXT:    store i8** [[ARGV]], i8*** [[ARGV_ADDR]], align 8
// CHECK3-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK3-NEXT:    [[ARRAYCTOR_END:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK3-NEXT:    br label [[ARRAYCTOR_LOOP:%.*]]
// CHECK3:       arrayctor.loop:
// CHECK3-NEXT:    [[ARRAYCTOR_CUR:%.*]] = phi %struct.S* [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[ARRAYCTOR_NEXT:%.*]], [[ARRAYCTOR_LOOP]] ]
// CHECK3-NEXT:    call void @_ZN1SC1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYCTOR_CUR]])
// CHECK3-NEXT:    [[ARRAYCTOR_NEXT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYCTOR_CUR]], i64 1
// CHECK3-NEXT:    [[ARRAYCTOR_DONE:%.*]] = icmp eq %struct.S* [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK3-NEXT:    br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK3:       arrayctor.cont:
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK3-NEXT:    [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
// CHECK3-NEXT:    [[TMP2:%.*]] = call i8* @llvm.stacksave()
// CHECK3-NEXT:    store i8* [[TMP2]], i8** [[SAVED_STACK]], align 8
// CHECK3-NEXT:    [[VLA:%.*]] = alloca i16, i64 [[TMP1]], align 16
// CHECK3-NEXT:    store i64 [[TMP1]], i64* [[__VLA_EXPR0]], align 8
// CHECK3-NEXT:    store i32 0, i32* [[I]], align 4
// CHECK3-NEXT:    br label [[FOR_COND:%.*]]
// CHECK3:       for.cond:
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32, i32* [[I]], align 4
// CHECK3-NEXT:    [[CMP:%.*]] = icmp slt i32 [[TMP3]], 5
// CHECK3-NEXT:    br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
// CHECK3:       for.body:
// CHECK3-NEXT:    [[TMP4:%.*]] = load i32, i32* [[A]], align 4
// CHECK3-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP4]] to i64
// CHECK3-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i16, i16* [[VLA]], i64 [[IDXPROM]]
// CHECK3-NEXT:    [[TMP5:%.*]] = load i16, i16* [[ARRAYIDX]], align 2
// CHECK3-NEXT:    [[CONV:%.*]] = sext i16 [[TMP5]] to i32
// CHECK3-NEXT:    [[TMP6:%.*]] = load i32, i32* [[A]], align 4
// CHECK3-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP6]], [[CONV]]
// CHECK3-NEXT:    store i32 [[ADD]], i32* [[A]], align 4
// CHECK3-NEXT:    br label [[FOR_INC:%.*]]
// CHECK3:       for.inc:
// CHECK3-NEXT:    [[TMP7:%.*]] = load i32, i32* [[I]], align 4
// CHECK3-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK3-NEXT:    store i32 [[INC]], i32* [[I]], align 4
// CHECK3-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP2:![0-9]+]]
// CHECK3:       for.end:
// CHECK3-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK3-NEXT:    [[TMP8:%.*]] = load i8*, i8** [[SAVED_STACK]], align 8
// CHECK3-NEXT:    call void @llvm.stackrestore(i8* [[TMP8]])
// CHECK3-NEXT:    [[ARRAY_BEGIN2:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK3-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAY_BEGIN2]], i64 5
// CHECK3-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK3:       arraydestroy.body:
// CHECK3-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP9]], [[FOR_END]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK3-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK3-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3:[0-9]+]]
// CHECK3-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN2]]
// CHECK3-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE3:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK3:       arraydestroy.done3:
// CHECK3-NEXT:    [[TMP10:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK3-NEXT:    ret i32 [[TMP10]]
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SC1Ev
// CHECK3-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1:[0-9]+]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK3-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    call void @_ZN1SC2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]])
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SD1Ev
// CHECK3-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK3-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    call void @_ZN1SD2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]]) #[[ATTR3]]
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SC2Ev
// CHECK3-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK3-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[A:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK3-NEXT:    store i32 0, i32* [[A]], align 4
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_ZN1SD2Ev
// CHECK3-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK3-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK3-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@main
// CHECK4-SAME: (i32 [[ARGC:%.*]], i8** [[ARGV:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[RETVAL:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[ARGV_ADDR:%.*]] = alloca i8**, align 8
// CHECK4-NEXT:    [[A:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    [[B:%.*]] = alloca float, align 4
// CHECK4-NEXT:    [[C:%.*]] = alloca [5 x %struct.S], align 16
// CHECK4-NEXT:    [[SAVED_STACK:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8
// CHECK4-NEXT:    [[DOTTASK_RED_:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    [[DOTTASK_RED_1:%.*]] = alloca i8*, align 8
// CHECK4-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK4-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK4-NEXT:    store i32 [[ARGC]], i32* [[ARGC_ADDR]], align 4
// CHECK4-NEXT:    store i8** [[ARGV]], i8*** [[ARGV_ADDR]], align 8
// CHECK4-NEXT:    [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK4-NEXT:    [[ARRAYCTOR_END:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[ARRAY_BEGIN]], i64 5
// CHECK4-NEXT:    br label [[ARRAYCTOR_LOOP:%.*]]
// CHECK4:       arrayctor.loop:
// CHECK4-NEXT:    [[ARRAYCTOR_CUR:%.*]] = phi %struct.S* [ [[ARRAY_BEGIN]], [[ENTRY:%.*]] ], [ [[ARRAYCTOR_NEXT:%.*]], [[ARRAYCTOR_LOOP]] ]
// CHECK4-NEXT:    call void @_ZN1SC1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYCTOR_CUR]])
// CHECK4-NEXT:    [[ARRAYCTOR_NEXT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYCTOR_CUR]], i64 1
// CHECK4-NEXT:    [[ARRAYCTOR_DONE:%.*]] = icmp eq %struct.S* [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK4-NEXT:    br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK4:       arrayctor.cont:
// CHECK4-NEXT:    [[TMP0:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK4-NEXT:    [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
// CHECK4-NEXT:    [[TMP2:%.*]] = call i8* @llvm.stacksave()
// CHECK4-NEXT:    store i8* [[TMP2]], i8** [[SAVED_STACK]], align 8
// CHECK4-NEXT:    [[VLA:%.*]] = alloca i16, i64 [[TMP1]], align 16
// CHECK4-NEXT:    store i64 [[TMP1]], i64* [[__VLA_EXPR0]], align 8
// CHECK4-NEXT:    store i32 0, i32* [[I]], align 4
// CHECK4-NEXT:    br label [[FOR_COND:%.*]]
// CHECK4:       for.cond:
// CHECK4-NEXT:    [[TMP3:%.*]] = load i32, i32* [[I]], align 4
// CHECK4-NEXT:    [[CMP:%.*]] = icmp slt i32 [[TMP3]], 5
// CHECK4-NEXT:    br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
// CHECK4:       for.body:
// CHECK4-NEXT:    [[TMP4:%.*]] = load i32, i32* [[A]], align 4
// CHECK4-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP4]] to i64
// CHECK4-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i16, i16* [[VLA]], i64 [[IDXPROM]]
// CHECK4-NEXT:    [[TMP5:%.*]] = load i16, i16* [[ARRAYIDX]], align 2
// CHECK4-NEXT:    [[CONV:%.*]] = sext i16 [[TMP5]] to i32
// CHECK4-NEXT:    [[TMP6:%.*]] = load i32, i32* [[A]], align 4
// CHECK4-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP6]], [[CONV]]
// CHECK4-NEXT:    store i32 [[ADD]], i32* [[A]], align 4
// CHECK4-NEXT:    br label [[FOR_INC:%.*]]
// CHECK4:       for.inc:
// CHECK4-NEXT:    [[TMP7:%.*]] = load i32, i32* [[I]], align 4
// CHECK4-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK4-NEXT:    store i32 [[INC]], i32* [[I]], align 4
// CHECK4-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP2:![0-9]+]]
// CHECK4:       for.end:
// CHECK4-NEXT:    store i32 0, i32* [[RETVAL]], align 4
// CHECK4-NEXT:    [[TMP8:%.*]] = load i8*, i8** [[SAVED_STACK]], align 8
// CHECK4-NEXT:    call void @llvm.stackrestore(i8* [[TMP8]])
// CHECK4-NEXT:    [[ARRAY_BEGIN2:%.*]] = getelementptr inbounds [5 x %struct.S], [5 x %struct.S]* [[C]], i32 0, i32 0
// CHECK4-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAY_BEGIN2]], i64 5
// CHECK4-NEXT:    br label [[ARRAYDESTROY_BODY:%.*]]
// CHECK4:       arraydestroy.body:
// CHECK4-NEXT:    [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi %struct.S* [ [[TMP9]], [[FOR_END]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK4-NEXT:    [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK4-NEXT:    call void @_ZN1SD1Ev(%struct.S* nonnull dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3:[0-9]+]]
// CHECK4-NEXT:    [[ARRAYDESTROY_DONE:%.*]] = icmp eq %struct.S* [[ARRAYDESTROY_ELEMENT]], [[ARRAY_BEGIN2]]
// CHECK4-NEXT:    br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE3:%.*]], label [[ARRAYDESTROY_BODY]]
// CHECK4:       arraydestroy.done3:
// CHECK4-NEXT:    [[TMP10:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK4-NEXT:    ret i32 [[TMP10]]
//
//
// CHECK4-LABEL: define {{[^@]+}}@_ZN1SC1Ev
// CHECK4-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1:[0-9]+]] align 2 {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK4-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    call void @_ZN1SC2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]])
// CHECK4-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@_ZN1SD1Ev
// CHECK4-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK4-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    call void @_ZN1SD2Ev(%struct.S* nonnull dereferenceable(4) [[THIS1]]) #[[ATTR3]]
// CHECK4-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@_ZN1SC2Ev
// CHECK4-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK4-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    [[A:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], %struct.S* [[THIS1]], i32 0, i32 0
// CHECK4-NEXT:    store i32 0, i32* [[A]], align 4
// CHECK4-NEXT:    ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@_ZN1SD2Ev
// CHECK4-SAME: (%struct.S* nonnull dereferenceable(4) [[THIS:%.*]]) unnamed_addr #[[ATTR1]] align 2 {
// CHECK4-NEXT:  entry:
// CHECK4-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK4-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK4-NEXT:    ret void
//