// 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 --implicit-check-not="{{__kmpc|__tgt}}"
// 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 --implicit-check-not="{{__kmpc|__tgt}}"
// expected-no-diagnostics
#ifndef HEADER
#define HEADER

int main(int argc, char **argv) {


#pragma omp parallel master taskloop priority(argc)
  for (int i = 0; i < 10; ++i)
    ;


#pragma omp parallel master taskloop nogroup grainsize(argc)
  for (int i = 0; i < 10; ++i)
    ;


  int i;
#pragma omp parallel master taskloop if(argc) shared(argc, argv) collapse(2) num_tasks(argc)
  for (i = 0; i < argc; ++i)
  for (int j = argc; j < argv[argc][argc]; ++j)
    ;
#pragma omp parallel master taskloop
  for (int i = 0; i < 10; ++i) {
#pragma omp cancel taskgroup
#pragma omp cancellation point taskgroup
  }
}

struct S {
  int a;
  S(int c) {

#pragma omp parallel master taskloop shared(c) num_tasks(4) final(c)
    for (a = 0; a < c; ++a)
      ;
  }
} s(1);


#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:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED2:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_4:%.*]] = alloca i8, align 1
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_5:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED7:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED10:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTBOUND_ZERO_ADDR:%.*]] = alloca i32, align 4
// 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:    [[TMP1:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP1]], i32* [[DOTCAPTURE_EXPR_]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_]], align 4
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED]] to i32*
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[CONV]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i64)* @.omp_outlined. to void (i32*, i32*, ...)*), i64 [[TMP3]])
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP4]], i32* [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[CONV3:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED2]] to i32*
// CHECK1-NEXT:    store i32 [[TMP5]], i32* [[CONV3]], align 4
// CHECK1-NEXT:    [[TMP6:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED2]], align 8
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i64)* @.omp_outlined..2 to void (i32*, i32*, ...)*), i64 [[TMP6]])
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[TMP7]], 0
// CHECK1-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[TOBOOL]] to i8
// CHECK1-NEXT:    store i8 [[FROMBOOL]], i8* [[DOTCAPTURE_EXPR_4]], align 1
// CHECK1-NEXT:    [[TMP8:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP8]], i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK1-NEXT:    [[TMP9:%.*]] = load i8, i8* [[DOTCAPTURE_EXPR_4]], align 1
// CHECK1-NEXT:    [[TOBOOL6:%.*]] = trunc i8 [[TMP9]] to i1
// CHECK1-NEXT:    [[CONV8:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED7]] to i8*
// CHECK1-NEXT:    [[FROMBOOL9:%.*]] = zext i1 [[TOBOOL6]] to i8
// CHECK1-NEXT:    store i8 [[FROMBOOL9]], i8* [[CONV8]], align 1
// CHECK1-NEXT:    [[TMP10:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED7]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK1-NEXT:    [[CONV11:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED10]] to i32*
// CHECK1-NEXT:    store i32 [[TMP11]], i32* [[CONV11]], align 4
// CHECK1-NEXT:    [[TMP12:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED10]], align 8
// CHECK1-NEXT:    [[TMP13:%.*]] = load i8, i8* [[DOTCAPTURE_EXPR_4]], align 1
// CHECK1-NEXT:    [[TOBOOL12:%.*]] = trunc i8 [[TMP13]] to i1
// CHECK1-NEXT:    br i1 [[TOBOOL12]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_ELSE:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i8***, i64, i64)* @.omp_outlined..5 to void (i32*, i32*, ...)*), i32* [[ARGC_ADDR]], i8*** [[ARGV_ADDR]], i64 [[TMP10]], i64 [[TMP12]])
// CHECK1-NEXT:    br label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.else:
// CHECK1-NEXT:    call void @__kmpc_serialized_parallel(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    store i32 [[TMP0]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTBOUND_ZERO_ADDR]], align 4
// CHECK1-NEXT:    call void @.omp_outlined..5(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTBOUND_ZERO_ADDR]], i32* [[ARGC_ADDR]], i8*** [[ARGV_ADDR]], i64 [[TMP10]], i64 [[TMP12]]) #[[ATTR2:[0-9]+]]
// CHECK1-NEXT:    call void @__kmpc_end_serialized_parallel(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* @.omp_outlined..8 to void (i32*, i32*, ...)*))
// CHECK1-NEXT:    [[TMP14:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK1-NEXT:    ret i32 [[TMP14]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[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:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 1
// 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 i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32*
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    [[TMP3:%.*]] = icmp ne i32 [[TMP2]], 0
// CHECK1-NEXT:    br i1 [[TMP3]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, i32* [[CONV]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 33, i64 80, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates*)* @.omp_task_entry. to i32 (i32, i8*)*))
// CHECK1-NEXT:    [[TMP6:%.*]] = bitcast i8* [[TMP5]] to %struct.kmp_task_t_with_privates*
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], %struct.kmp_task_t_with_privates* [[TMP6]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 4
// CHECK1-NEXT:    [[TMP9:%.*]] = bitcast %union.kmp_cmplrdata_t* [[TMP8]] to i32*
// CHECK1-NEXT:    store i32 [[TMP4]], i32* [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 6
// CHECK1-NEXT:    store i64 9, i64* [[TMP11]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, i64* [[TMP12]], align 8
// CHECK1-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP14:%.*]] = bitcast i8** [[TMP13]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP14]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP15:%.*]] = load i64, i64* [[TMP12]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i8* [[TMP5]], i32 1, i64* [[TMP10]], i64* [[TMP11]], i64 [[TMP15]], i32 1, i32 0, i64 0, i8* null)
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry.
// CHECK1-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates* noalias [[TMP1:%.*]]) #[[ATTR4:[0-9]+]] {
// 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:    [[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:%.*]] = bitcast %struct.kmp_task_t_with_privates* [[TMP3]] to i8*
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META3:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !14
// CHECK1-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !14
// CHECK1-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    store %struct.anon* [[TMP8]], %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP20:%.*]] = load %struct.anon*, %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[TMP21:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP21]] to i32
// CHECK1-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP22:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK1-NEXT:    [[CONV1_I:%.*]] = sext i32 [[TMP22]] to i64
// CHECK1-NEXT:    [[TMP23:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !14
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV1_I]], [[TMP23]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__1_EXIT:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP24:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK1-NEXT:    store i32 [[TMP24]], i32* [[I_I]], align 4, !noalias !14
// CHECK1-NEXT:    [[TMP25:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK1-NEXT:    [[ADD2_I:%.*]] = add nsw i32 [[TMP25]], 1
// CHECK1-NEXT:    store i32 [[ADD2_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK1:       .omp_outlined..1.exit:
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..2
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_0:%.*]], align 1
// 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 i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32*
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    [[TMP3:%.*]] = icmp ne i32 [[TMP2]], 0
// CHECK1-NEXT:    br i1 [[TMP3]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, i32* [[CONV]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 1, i64 80, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.1*)* @.omp_task_entry..4 to i32 (i32, i8*)*))
// CHECK1-NEXT:    [[TMP6:%.*]] = bitcast i8* [[TMP5]] to %struct.kmp_task_t_with_privates.1*
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_1:%.*]], %struct.kmp_task_t_with_privates.1* [[TMP6]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, i64* [[TMP8]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 6
// CHECK1-NEXT:    store i64 9, i64* [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP12:%.*]] = bitcast i8** [[TMP11]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP12]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP13:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP14:%.*]] = zext i32 [[TMP4]] to i64
// CHECK1-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i8* [[TMP5]], i32 1, i64* [[TMP8]], i64* [[TMP9]], i64 [[TMP13]], i32 1, i32 1, i64 [[TMP14]], i8* null)
// CHECK1-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry..4
// CHECK1-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.1* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.0*, 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.1*, align 8
// CHECK1-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    store %struct.kmp_task_t_with_privates.1* [[TMP1]], %struct.kmp_task_t_with_privates.1** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.1*, %struct.kmp_task_t_with_privates.1** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_1:%.*]], %struct.kmp_task_t_with_privates.1* [[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.0*
// CHECK1-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.1* [[TMP3]] to i8*
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META17:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META20:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META22:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META24:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META26:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !28
// CHECK1-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !28
// CHECK1-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    store %struct.anon.0* [[TMP8]], %struct.anon.0** [[__CONTEXT_ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    [[TMP20:%.*]] = load %struct.anon.0*, %struct.anon.0** [[__CONTEXT_ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    [[TMP21:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP21]] to i32
// CHECK1-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP22:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK1-NEXT:    [[CONV1_I:%.*]] = sext i32 [[TMP22]] to i64
// CHECK1-NEXT:    [[TMP23:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !28
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV1_I]], [[TMP23]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__3_EXIT:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP24:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK1-NEXT:    store i32 [[TMP24]], i32* [[I_I]], align 4, !noalias !28
// CHECK1-NEXT:    [[TMP25:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK1-NEXT:    [[ADD2_I:%.*]] = add nsw i32 [[TMP25]], 1
// CHECK1-NEXT:    store i32 [[ADD2_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK1:       .omp_outlined..3.exit:
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..5
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i8*** nonnull align 8 dereferenceable(8) [[ARGV:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]], i64 [[DOTCAPTURE_EXPR_1:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[ARGV_ADDR:%.*]] = alloca i8***, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR2:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_2:%.*]], align 8
// CHECK1-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[_TMP4:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_5:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_6:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_7:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_11:%.*]] = alloca i64, align 8
// 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* [[ARGC]], i32** [[ARGC_ADDR]], align 8
// CHECK1-NEXT:    store i8*** [[ARGV]], i8**** [[ARGV_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTCAPTURE_EXPR_1]], i64* [[DOTCAPTURE_EXPR__ADDR2]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[ARGC_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i8***, i8**** [[ARGV_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i8*
// CHECK1-NEXT:    [[CONV3:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR2]] to i32*
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    [[TMP5:%.*]] = icmp ne i32 [[TMP4]], 0
// CHECK1-NEXT:    br i1 [[TMP5]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT:    store i32* [[TMP0]], i32** [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[AGG_CAPTURED]], i32 0, i32 1
// CHECK1-NEXT:    store i8*** [[TMP1]], i8**** [[TMP7]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = load i32, i32* [[CONV3]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    store i32 [[TMP9]], i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK1-NEXT:    [[TMP10:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    store i32 [[TMP10]], i32* [[DOTCAPTURE_EXPR_6]], align 4
// CHECK1-NEXT:    [[TMP11:%.*]] = load i8**, i8*** [[TMP1]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP12]] to i64
// CHECK1-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8*, i8** [[TMP11]], i64 [[IDXPROM]]
// CHECK1-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[ARRAYIDX]], align 8
// CHECK1-NEXT:    [[TMP14:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[TMP14]] to i64
// CHECK1-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds i8, i8* [[TMP13]], i64 [[IDXPROM8]]
// CHECK1-NEXT:    [[TMP15:%.*]] = load i8, i8* [[ARRAYIDX9]], align 1
// CHECK1-NEXT:    [[CONV10:%.*]] = sext i8 [[TMP15]] to i32
// CHECK1-NEXT:    store i32 [[CONV10]], i32* [[DOTCAPTURE_EXPR_7]], align 4
// CHECK1-NEXT:    [[TMP16:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK1-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP16]], 0
// CHECK1-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK1-NEXT:    [[CONV12:%.*]] = sext i32 [[DIV]] to i64
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_7]], align 4
// CHECK1-NEXT:    [[TMP18:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_6]], align 4
// CHECK1-NEXT:    [[SUB13:%.*]] = sub i32 [[TMP17]], [[TMP18]]
// CHECK1-NEXT:    [[SUB14:%.*]] = sub i32 [[SUB13]], 1
// CHECK1-NEXT:    [[ADD:%.*]] = add i32 [[SUB14]], 1
// CHECK1-NEXT:    [[DIV15:%.*]] = udiv i32 [[ADD]], 1
// CHECK1-NEXT:    [[CONV16:%.*]] = zext i32 [[DIV15]] to i64
// CHECK1-NEXT:    [[MUL:%.*]] = mul nsw i64 [[CONV12]], [[CONV16]]
// CHECK1-NEXT:    [[SUB17:%.*]] = sub nsw i64 [[MUL]], 1
// CHECK1-NEXT:    store i64 [[SUB17]], i64* [[DOTCAPTURE_EXPR_11]], align 8
// CHECK1-NEXT:    [[TMP19:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i32 1, i64 80, i64 16, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.3*)* @.omp_task_entry..7 to i32 (i32, i8*)*))
// CHECK1-NEXT:    [[TMP20:%.*]] = bitcast i8* [[TMP19]] to %struct.kmp_task_t_with_privates.3*
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_3:%.*]], %struct.kmp_task_t_with_privates.3* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP23:%.*]] = load i8*, i8** [[TMP22]], align 8
// CHECK1-NEXT:    [[TMP24:%.*]] = bitcast %struct.anon.2* [[AGG_CAPTURED]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP23]], i8* align 8 [[TMP24]], i64 16, i1 false)
// CHECK1-NEXT:    [[TMP25:%.*]] = load i8, i8* [[CONV]], align 8
// CHECK1-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[TMP25]] to i1
// CHECK1-NEXT:    [[TMP26:%.*]] = sext i1 [[TOBOOL]] to i32
// CHECK1-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, i64* [[TMP27]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP29:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR_11]], align 8
// CHECK1-NEXT:    store i64 [[TMP29]], i64* [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, i64* [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP31:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP32:%.*]] = bitcast i8** [[TMP31]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP32]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP33:%.*]] = load i64, i64* [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP34:%.*]] = zext i32 [[TMP8]] to i64
// CHECK1-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i8* [[TMP19]], i32 [[TMP26]], i64* [[TMP27]], i64* [[TMP28]], i64 [[TMP33]], i32 1, i32 2, i64 [[TMP34]], i8* null)
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry..7
// CHECK1-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.3* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.2*, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_2_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_3_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_6_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[J_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[I14_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[J15_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca %struct.kmp_task_t_with_privates.3*, align 8
// CHECK1-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    store %struct.kmp_task_t_with_privates.3* [[TMP1]], %struct.kmp_task_t_with_privates.3** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.3*, %struct.kmp_task_t_with_privates.3** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_3:%.*]], %struct.kmp_task_t_with_privates.3* [[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.2*
// CHECK1-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.3* [[TMP3]] to i8*
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META29:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META32:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META34:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META36:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META38:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !40
// CHECK1-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !40
// CHECK1-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store %struct.anon.2* [[TMP8]], %struct.anon.2** [[__CONTEXT_ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP20:%.*]] = load %struct.anon.2*, %struct.anon.2** [[__CONTEXT_ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_ANON_2:%.*]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP22:%.*]] = load i32*, i32** [[TMP21]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = load i32, i32* [[TMP22]], align 4
// CHECK1-NEXT:    store i32 [[TMP23]], i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
// CHECK1-NEXT:    store i32 [[TMP26]], i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP28:%.*]] = load i8***, i8**** [[TMP27]], align 8
// CHECK1-NEXT:    [[TMP29:%.*]] = load i8**, i8*** [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP31:%.*]] = load i32*, i32** [[TMP30]], align 8
// CHECK1-NEXT:    [[TMP32:%.*]] = load i32, i32* [[TMP31]], align 4
// CHECK1-NEXT:    [[IDXPROM_I:%.*]] = sext i32 [[TMP32]] to i64
// CHECK1-NEXT:    [[ARRAYIDX_I:%.*]] = getelementptr inbounds i8*, i8** [[TMP29]], i64 [[IDXPROM_I]]
// CHECK1-NEXT:    [[TMP33:%.*]] = load i8*, i8** [[ARRAYIDX_I]], align 8
// CHECK1-NEXT:    [[TMP34:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP35:%.*]] = load i32*, i32** [[TMP34]], align 8
// CHECK1-NEXT:    [[TMP36:%.*]] = load i32, i32* [[TMP35]], align 4
// CHECK1-NEXT:    [[IDXPROM4_I:%.*]] = sext i32 [[TMP36]] to i64
// CHECK1-NEXT:    [[ARRAYIDX5_I:%.*]] = getelementptr inbounds i8, i8* [[TMP33]], i64 [[IDXPROM4_I]]
// CHECK1-NEXT:    [[TMP37:%.*]] = load i8, i8* [[ARRAYIDX5_I]], align 1
// CHECK1-NEXT:    [[CONV_I:%.*]] = sext i8 [[TMP37]] to i32
// CHECK1-NEXT:    store i32 [[CONV_I]], i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP38:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !40
// CHECK1-NEXT:    [[CONV7_I:%.*]] = sext i32 [[TMP38]] to i64
// CHECK1-NEXT:    [[TMP39:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP40:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[SUB8_I:%.*]] = sub i32 [[TMP39]], [[TMP40]]
// CHECK1-NEXT:    [[SUB9_I:%.*]] = sub i32 [[SUB8_I]], 1
// CHECK1-NEXT:    [[CONV11_I:%.*]] = zext i32 [[SUB8_I]] to i64
// CHECK1-NEXT:    [[MUL_I:%.*]] = mul nsw i64 [[CONV7_I]], [[CONV11_I]]
// CHECK1-NEXT:    [[SUB12_I:%.*]] = sub nsw i64 [[MUL_I]], 1
// CHECK1-NEXT:    store i64 [[SUB12_I]], i64* [[DOTCAPTURE_EXPR_6_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i32 0, i32* [[I_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP41:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    store i32 [[TMP41]], i32* [[J_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP42:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !40
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp slt i32 0, [[TMP42]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[LAND_LHS_TRUE_I:%.*]], label [[DOTOMP_OUTLINED__6_EXIT:%.*]]
// CHECK1:       land.lhs.true.i:
// CHECK1-NEXT:    [[TMP43:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP44:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[CMP13_I:%.*]] = icmp slt i32 [[TMP43]], [[TMP44]]
// CHECK1-NEXT:    br i1 [[CMP13_I]], label [[TASKLOOP_IF_THEN_I:%.*]], label [[DOTOMP_OUTLINED__6_EXIT]]
// CHECK1:       taskloop.if.then.i:
// CHECK1-NEXT:    [[TMP45:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    store i64 [[TMP45]], i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP46:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP47:%.*]] = load i32*, i32** [[TMP46]], align 8
// CHECK1-NEXT:    [[TMP48:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP49:%.*]] = load i8***, i8**** [[TMP48]], align 8
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP50:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP51:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[CMP16_I:%.*]] = icmp ule i64 [[TMP50]], [[TMP51]]
// CHECK1-NEXT:    br i1 [[CMP16_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[OMP_INNER_FOR_END_I:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP52:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP53:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP54:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[SUB17_I:%.*]] = sub i32 [[TMP53]], [[TMP54]]
// CHECK1-NEXT:    [[SUB18_I:%.*]] = sub i32 [[SUB17_I]], 1
// CHECK1-NEXT:    [[CONV22_I:%.*]] = zext i32 [[SUB17_I]] to i64
// CHECK1-NEXT:    [[DIV23_I:%.*]] = sdiv i64 [[TMP52]], [[CONV22_I]]
// CHECK1-NEXT:    [[CONV26_I:%.*]] = trunc i64 [[DIV23_I]] to i32
// CHECK1-NEXT:    store i32 [[CONV26_I]], i32* [[I14_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP55:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[CONV27_I:%.*]] = sext i32 [[TMP55]] to i64
// CHECK1-NEXT:    [[TMP56:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP57:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[TMP58:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP59:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[SUB28_I:%.*]] = sub i32 [[TMP58]], [[TMP59]]
// CHECK1-NEXT:    [[SUB29_I:%.*]] = sub i32 [[SUB28_I]], 1
// CHECK1-NEXT:    [[CONV33_I:%.*]] = zext i32 [[SUB28_I]] to i64
// CHECK1-NEXT:    [[DIV34_I:%.*]] = sdiv i64 [[TMP57]], [[CONV33_I]]
// CHECK1-NEXT:    [[TMP60:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP61:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[SUB35_I:%.*]] = sub i32 [[TMP60]], [[TMP61]]
// CHECK1-NEXT:    [[SUB36_I:%.*]] = sub i32 [[SUB35_I]], 1
// CHECK1-NEXT:    [[CONV40_I:%.*]] = zext i32 [[SUB35_I]] to i64
// CHECK1-NEXT:    [[MUL41_I:%.*]] = mul nsw i64 [[DIV34_I]], [[CONV40_I]]
// CHECK1-NEXT:    [[SUB42_I:%.*]] = sub nsw i64 [[TMP56]], [[MUL41_I]]
// CHECK1-NEXT:    [[ADD44_I:%.*]] = add nsw i64 [[CONV27_I]], [[SUB42_I]]
// CHECK1-NEXT:    [[CONV45_I:%.*]] = trunc i64 [[ADD44_I]] to i32
// CHECK1-NEXT:    store i32 [[CONV45_I]], i32* [[J15_I]], align 4, !noalias !40
// CHECK1-NEXT:    [[TMP62:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    [[ADD46_I:%.*]] = add nsw i64 [[TMP62]], 1
// CHECK1-NEXT:    store i64 [[ADD46_I]], i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK1:       omp.inner.for.end.i:
// CHECK1-NEXT:    br label [[DOTOMP_OUTLINED__6_EXIT]]
// CHECK1:       .omp_outlined..6.exit:
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..8
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_4:%.*]], align 1
// 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:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    [[TMP3:%.*]] = icmp ne i32 [[TMP2]], 0
// CHECK1-NEXT:    br i1 [[TMP3]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    [[TMP4:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 1, i64 80, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.5*)* @.omp_task_entry..10 to i32 (i32, i8*)*))
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to %struct.kmp_task_t_with_privates.5*
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_5:%.*]], %struct.kmp_task_t_with_privates.5* [[TMP5]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, i64* [[TMP7]], align 8
// CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 6
// CHECK1-NEXT:    store i64 9, i64* [[TMP8]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, i64* [[TMP9]], align 8
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP11:%.*]] = bitcast i8** [[TMP10]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP11]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP12:%.*]] = load i64, i64* [[TMP9]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i8* [[TMP4]], i32 1, i64* [[TMP7]], i64* [[TMP8]], i64 [[TMP12]], i32 1, i32 0, i64 0, i8* null)
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry..10
// CHECK1-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.5* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.4*, align 8
// CHECK1-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[CLEANUP_DEST_SLOT_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca %struct.kmp_task_t_with_privates.5*, align 8
// CHECK1-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    store %struct.kmp_task_t_with_privates.5* [[TMP1]], %struct.kmp_task_t_with_privates.5** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.5*, %struct.kmp_task_t_with_privates.5** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_5:%.*]], %struct.kmp_task_t_with_privates.5* [[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.4*
// CHECK1-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.5* [[TMP3]] to i8*
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META41:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META44:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META46:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META48:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META50:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !52
// CHECK1-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !52
// CHECK1-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    store %struct.anon.4* [[TMP8]], %struct.anon.4** [[__CONTEXT_ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    [[TMP20:%.*]] = load %struct.anon.4*, %struct.anon.4** [[__CONTEXT_ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    [[TMP21:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP21]] to i32
// CHECK1-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP22:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK1-NEXT:    [[CONV1_I:%.*]] = sext i32 [[TMP22]] to i64
// CHECK1-NEXT:    [[TMP23:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !52
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV1_I]], [[TMP23]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[OMP_INNER_FOR_END_I:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP24:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK1-NEXT:    store i32 [[TMP24]], i32* [[I_I]], align 4, !noalias !52
// CHECK1-NEXT:    [[TMP25:%.*]] = load i32, i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !52
// CHECK1-NEXT:    [[TMP26:%.*]] = call i32 @__kmpc_cancel(%struct.ident_t* @[[GLOB1]], i32 [[TMP25]], i32 4) #[[ATTR2]]
// CHECK1-NEXT:    [[TMP27:%.*]] = icmp ne i32 [[TMP26]], 0
// CHECK1-NEXT:    br i1 [[TMP27]], label [[DOTCANCEL_EXIT_I:%.*]], label [[DOTCANCEL_CONTINUE_I:%.*]]
// CHECK1:       .cancel.exit.i:
// CHECK1-NEXT:    store i32 1, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK1-NEXT:    br label [[DOTOMP_OUTLINED__9_EXIT:%.*]]
// CHECK1:       .cancel.continue.i:
// CHECK1-NEXT:    [[TMP28:%.*]] = load i32, i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !52
// CHECK1-NEXT:    [[TMP29:%.*]] = call i32 @__kmpc_cancellationpoint(%struct.ident_t* @[[GLOB1]], i32 [[TMP28]], i32 4) #[[ATTR2]]
// CHECK1-NEXT:    [[TMP30:%.*]] = icmp ne i32 [[TMP29]], 0
// CHECK1-NEXT:    br i1 [[TMP30]], label [[DOTCANCEL_EXIT2_I:%.*]], label [[DOTCANCEL_CONTINUE3_I:%.*]]
// CHECK1:       .cancel.exit2.i:
// CHECK1-NEXT:    store i32 1, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK1-NEXT:    br label [[DOTOMP_OUTLINED__9_EXIT]]
// CHECK1:       .cancel.continue3.i:
// CHECK1-NEXT:    [[TMP31:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK1-NEXT:    [[ADD4_I:%.*]] = add nsw i32 [[TMP31]], 1
// CHECK1-NEXT:    store i32 [[ADD4_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK1:       omp.inner.for.end.i:
// CHECK1-NEXT:    store i32 0, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK1-NEXT:    br label [[DOTOMP_OUTLINED__9_EXIT]]
// CHECK1:       .omp_outlined..9.exit:
// CHECK1-NEXT:    [[CLEANUP_DEST_I:%.*]] = load i32, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@__cxx_global_var_init
// CHECK1-SAME: () #[[ATTR7:[0-9]+]] section "__TEXT,__StaticInit,regular,pure_instructions" {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    call void @_ZN1SC1Ei(%struct.S* nonnull align 4 dereferenceable(4) @s, i32 1)
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC1Ei
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[C:%.*]]) unnamed_addr #[[ATTR8:[0-9]+]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store i32 [[C]], i32* [[C_ADDR]], align 4
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32, i32* [[C_ADDR]], align 4
// CHECK1-NEXT:    call void @_ZN1SC2Ei(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]], i32 [[TMP0]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_ZN1SC2Ei
// CHECK1-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[C:%.*]]) unnamed_addr #[[ATTR8]] align 2 {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i8, align 1
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store i32 [[C]], i32* [[C_ADDR]], align 4
// CHECK1-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32, i32* [[C_ADDR]], align 4
// CHECK1-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[TMP0]], 0
// CHECK1-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[TOBOOL]] to i8
// CHECK1-NEXT:    store i8 [[FROMBOOL]], i8* [[DOTCAPTURE_EXPR_]], align 1
// CHECK1-NEXT:    [[TMP1:%.*]] = load i8, i8* [[DOTCAPTURE_EXPR_]], align 1
// CHECK1-NEXT:    [[TOBOOL2:%.*]] = trunc i8 [[TMP1]] to i1
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED]] to i8*
// CHECK1-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[TOBOOL2]] to i8
// CHECK1-NEXT:    store i8 [[FROMBOOL3]], i8* [[CONV]], align 1
// CHECK1-NEXT:    [[TMP2:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, %struct.S*, i32*, i64)* @.omp_outlined..11 to void (i32*, i32*, ...)*), %struct.S* [[THIS1]], i32* [[C_ADDR]], i64 [[TMP2]])
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..11
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], %struct.S* [[THIS:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_6:%.*]], align 8
// CHECK1-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[_TMP1:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_3:%.*]] = 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 %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i8*
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    [[TMP5:%.*]] = icmp ne i32 [[TMP4]], 0
// CHECK1-NEXT:    br i1 [[TMP5]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1:       omp_if.then:
// CHECK1-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT:    store %struct.S* [[TMP0]], %struct.S** [[TMP6]], align 8
// CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[AGG_CAPTURED]], i32 0, i32 1
// CHECK1-NEXT:    store i32* [[TMP1]], i32** [[TMP7]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    [[TMP8:%.*]] = load i8, i8* [[CONV]], align 8
// CHECK1-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[TMP8]] to i1
// CHECK1-NEXT:    store i32* [[TMP]], i32** [[_TMP1]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP1]], align 4
// CHECK1-NEXT:    store i32 [[TMP9]], i32* [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    [[TMP10:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP10]], 0
// CHECK1-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK1-NEXT:    [[SUB4:%.*]] = sub nsw i32 [[DIV]], 1
// CHECK1-NEXT:    store i32 [[SUB4]], i32* [[DOTCAPTURE_EXPR_3]], align 4
// CHECK1-NEXT:    [[TMP11:%.*]] = select i1 [[TOBOOL]], i32 2, i32 0
// CHECK1-NEXT:    [[TMP12:%.*]] = or i32 [[TMP11]], 1
// CHECK1-NEXT:    [[TMP13:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i32 [[TMP12]], i64 80, i64 16, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.7*)* @.omp_task_entry..13 to i32 (i32, i8*)*))
// CHECK1-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to %struct.kmp_task_t_with_privates.7*
// CHECK1-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_7:%.*]], %struct.kmp_task_t_with_privates.7* [[TMP14]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP17:%.*]] = load i8*, i8** [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = bitcast %struct.anon.6* [[AGG_CAPTURED]] to i8*
// CHECK1-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP17]], i8* align 8 [[TMP18]], i64 16, i1 false)
// CHECK1-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 5
// CHECK1-NEXT:    store i64 0, i64* [[TMP19]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 6
// CHECK1-NEXT:    [[TMP21:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3]], align 4
// CHECK1-NEXT:    [[CONV5:%.*]] = sext i32 [[TMP21]] to i64
// CHECK1-NEXT:    store i64 [[CONV5]], i64* [[TMP20]], align 8
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 7
// CHECK1-NEXT:    store i64 1, i64* [[TMP22]], align 8
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP24:%.*]] = bitcast i8** [[TMP23]] to i8*
// CHECK1-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP24]], i8 0, i64 8, i1 false)
// CHECK1-NEXT:    [[TMP25:%.*]] = load i64, i64* [[TMP22]], align 8
// CHECK1-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i8* [[TMP13]], i32 1, i64* [[TMP19]], i64* [[TMP20]], i64 [[TMP25]], i32 1, i32 2, i64 4, i8* null)
// CHECK1-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK1-NEXT:    br label [[OMP_IF_END]]
// CHECK1:       omp_if.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_task_entry..13
// CHECK1-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.7* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.6*, align 8
// CHECK1-NEXT:    [[TMP_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TMP1_I:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_2_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[A_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TMP4_I:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[A5_I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TMP6_I:%.*]] = alloca i32*, align 8
// 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.7*, align 8
// CHECK1-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    store %struct.kmp_task_t_with_privates.7* [[TMP1]], %struct.kmp_task_t_with_privates.7** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.7*, %struct.kmp_task_t_with_privates.7** [[DOTADDR1]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_7:%.*]], %struct.kmp_task_t_with_privates.7* [[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.6*
// CHECK1-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.7* [[TMP3]] to i8*
// CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK1-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK1-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META53:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META56:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META58:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META60:![0-9]+]])
// CHECK1-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META62:![0-9]+]])
// CHECK1-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !64
// CHECK1-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !64
// CHECK1-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    store %struct.anon.6* [[TMP8]], %struct.anon.6** [[__CONTEXT_ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[TMP20:%.*]] = load %struct.anon.6*, %struct.anon.6** [[__CONTEXT_ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_ANON_6:%.*]], %struct.anon.6* [[TMP20]], i32 0, i32 0
// CHECK1-NEXT:    [[TMP22:%.*]] = load %struct.S*, %struct.S** [[TMP21]], align 8
// CHECK1-NEXT:    store i32* [[TMP_I]], i32** [[TMP1_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[TMP20]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP24:%.*]] = load i32*, i32** [[TMP23]], align 8
// CHECK1-NEXT:    [[TMP25:%.*]] = load i32, i32* [[TMP24]], align 4
// CHECK1-NEXT:    store i32 [[TMP25]], i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !64
// CHECK1-NEXT:    [[TMP26:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !64
// CHECK1-NEXT:    [[SUB3_I:%.*]] = sub nsw i32 [[TMP26]], 1
// CHECK1-NEXT:    store i32 [[SUB3_I]], i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !64
// CHECK1-NEXT:    store i32* [[A_I]], i32** [[TMP4_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[TMP27:%.*]] = load i32*, i32** [[TMP4_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i32 0, i32* [[TMP27]], align 4
// CHECK1-NEXT:    [[TMP28:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !64
// CHECK1-NEXT:    [[CMP_I:%.*]] = icmp slt i32 0, [[TMP28]]
// CHECK1-NEXT:    br i1 [[CMP_I]], label [[TASKLOOP_IF_THEN_I:%.*]], label [[DOTOMP_OUTLINED__12_EXIT:%.*]]
// CHECK1:       taskloop.if.then.i:
// CHECK1-NEXT:    store i32* [[A5_I]], i32** [[TMP6_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[TMP29:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP29]] to i32
// CHECK1-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK1-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[TMP20]], i32 0, i32 1
// CHECK1-NEXT:    [[TMP31:%.*]] = load i32*, i32** [[TMP30]], align 8
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1:       omp.inner.for.cond.i:
// CHECK1-NEXT:    [[TMP32:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK1-NEXT:    [[CONV7_I:%.*]] = sext i32 [[TMP32]] to i64
// CHECK1-NEXT:    [[TMP33:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !64
// CHECK1-NEXT:    [[CMP8_I:%.*]] = icmp ule i64 [[CONV7_I]], [[TMP33]]
// CHECK1-NEXT:    br i1 [[CMP8_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[OMP_INNER_FOR_END_I:%.*]]
// CHECK1:       omp.inner.for.body.i:
// CHECK1-NEXT:    [[TMP34:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK1-NEXT:    [[TMP35:%.*]] = load i32*, i32** [[TMP6_I]], align 8, !noalias !64
// CHECK1-NEXT:    store i32 [[TMP34]], i32* [[TMP35]], align 4
// CHECK1-NEXT:    [[TMP36:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK1-NEXT:    [[ADD9_I:%.*]] = add nsw i32 [[TMP36]], 1
// CHECK1-NEXT:    store i32 [[ADD9_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK1:       omp.inner.for.end.i:
// CHECK1-NEXT:    br label [[DOTOMP_OUTLINED__12_EXIT]]
// CHECK1:       .omp_outlined..12.exit:
// CHECK1-NEXT:    ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@_GLOBAL__sub_I_parallel_master_taskloop_codegen.cpp
// CHECK1-SAME: () #[[ATTR7]] section "__TEXT,__StaticInit,regular,pure_instructions" {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    call void @__cxx_global_var_init()
// 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:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED2:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_4:%.*]] = alloca i8, align 1
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_5:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED7:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED10:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTBOUND_ZERO_ADDR:%.*]] = alloca i32, align 4
// 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:    [[TMP1:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP1]], i32* [[DOTCAPTURE_EXPR_]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED]] to i32*
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[CONV]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i64)* @.omp_outlined. to void (i32*, i32*, ...)*), i64 [[TMP3]])
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP4]], i32* [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[TMP5:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[CONV3:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED2]] to i32*
// CHECK2-NEXT:    store i32 [[TMP5]], i32* [[CONV3]], align 4
// CHECK2-NEXT:    [[TMP6:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED2]], align 8
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i64)* @.omp_outlined..2 to void (i32*, i32*, ...)*), i64 [[TMP6]])
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[TMP7]], 0
// CHECK2-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[TOBOOL]] to i8
// CHECK2-NEXT:    store i8 [[FROMBOOL]], i8* [[DOTCAPTURE_EXPR_4]], align 1
// CHECK2-NEXT:    [[TMP8:%.*]] = load i32, i32* [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP8]], i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK2-NEXT:    [[TMP9:%.*]] = load i8, i8* [[DOTCAPTURE_EXPR_4]], align 1
// CHECK2-NEXT:    [[TOBOOL6:%.*]] = trunc i8 [[TMP9]] to i1
// CHECK2-NEXT:    [[CONV8:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED7]] to i8*
// CHECK2-NEXT:    [[FROMBOOL9:%.*]] = zext i1 [[TOBOOL6]] to i8
// CHECK2-NEXT:    store i8 [[FROMBOOL9]], i8* [[CONV8]], align 1
// CHECK2-NEXT:    [[TMP10:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED7]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK2-NEXT:    [[CONV11:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED10]] to i32*
// CHECK2-NEXT:    store i32 [[TMP11]], i32* [[CONV11]], align 4
// CHECK2-NEXT:    [[TMP12:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED10]], align 8
// CHECK2-NEXT:    [[TMP13:%.*]] = load i8, i8* [[DOTCAPTURE_EXPR_4]], align 1
// CHECK2-NEXT:    [[TOBOOL12:%.*]] = trunc i8 [[TMP13]] to i1
// CHECK2-NEXT:    br i1 [[TOBOOL12]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_ELSE:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i8***, i64, i64)* @.omp_outlined..5 to void (i32*, i32*, ...)*), i32* [[ARGC_ADDR]], i8*** [[ARGV_ADDR]], i64 [[TMP10]], i64 [[TMP12]])
// CHECK2-NEXT:    br label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.else:
// CHECK2-NEXT:    call void @__kmpc_serialized_parallel(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK2-NEXT:    store i32 [[TMP0]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTBOUND_ZERO_ADDR]], align 4
// CHECK2-NEXT:    call void @.omp_outlined..5(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTBOUND_ZERO_ADDR]], i32* [[ARGC_ADDR]], i8*** [[ARGV_ADDR]], i64 [[TMP10]], i64 [[TMP12]]) #[[ATTR2:[0-9]+]]
// CHECK2-NEXT:    call void @__kmpc_end_serialized_parallel(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* @.omp_outlined..8 to void (i32*, i32*, ...)*))
// CHECK2-NEXT:    [[TMP14:%.*]] = load i32, i32* [[RETVAL]], align 4
// CHECK2-NEXT:    ret i32 [[TMP14]]
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[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:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]], align 1
// 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 i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32*
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    [[TMP3:%.*]] = icmp ne i32 [[TMP2]], 0
// CHECK2-NEXT:    br i1 [[TMP3]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, i32* [[CONV]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 33, i64 80, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates*)* @.omp_task_entry. to i32 (i32, i8*)*))
// CHECK2-NEXT:    [[TMP6:%.*]] = bitcast i8* [[TMP5]] to %struct.kmp_task_t_with_privates*
// CHECK2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], %struct.kmp_task_t_with_privates* [[TMP6]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 4
// CHECK2-NEXT:    [[TMP9:%.*]] = bitcast %union.kmp_cmplrdata_t* [[TMP8]] to i32*
// CHECK2-NEXT:    store i32 [[TMP4]], i32* [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 5
// CHECK2-NEXT:    store i64 0, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 6
// CHECK2-NEXT:    store i64 9, i64* [[TMP11]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 7
// CHECK2-NEXT:    store i64 1, i64* [[TMP12]], align 8
// CHECK2-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP14:%.*]] = bitcast i8** [[TMP13]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP14]], i8 0, i64 8, i1 false)
// CHECK2-NEXT:    [[TMP15:%.*]] = load i64, i64* [[TMP12]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i8* [[TMP5]], i32 1, i64* [[TMP10]], i64* [[TMP11]], i64 [[TMP15]], i32 1, i32 0, i64 0, i8* null)
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_entry.
// CHECK2-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates* noalias [[TMP1:%.*]]) #[[ATTR4:[0-9]+]] {
// 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:    [[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:%.*]] = bitcast %struct.kmp_task_t_with_privates* [[TMP3]] to i8*
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK2-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META3:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !14
// CHECK2-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !14
// CHECK2-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    store %struct.anon* [[TMP8]], %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    [[TMP20:%.*]] = load %struct.anon*, %struct.anon** [[__CONTEXT_ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    [[TMP21:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP21]] to i32
// CHECK2-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK2:       omp.inner.for.cond.i:
// CHECK2-NEXT:    [[TMP22:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK2-NEXT:    [[CONV1_I:%.*]] = sext i32 [[TMP22]] to i64
// CHECK2-NEXT:    [[TMP23:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !14
// CHECK2-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV1_I]], [[TMP23]]
// CHECK2-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__1_EXIT:%.*]]
// CHECK2:       omp.inner.for.body.i:
// CHECK2-NEXT:    [[TMP24:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK2-NEXT:    store i32 [[TMP24]], i32* [[I_I]], align 4, !noalias !14
// CHECK2-NEXT:    [[TMP25:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK2-NEXT:    [[ADD2_I:%.*]] = add nsw i32 [[TMP25]], 1
// CHECK2-NEXT:    store i32 [[ADD2_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !14
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK2:       .omp_outlined..1.exit:
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined..2
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_0:%.*]], align 1
// 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 i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32*
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    [[TMP3:%.*]] = icmp ne i32 [[TMP2]], 0
// CHECK2-NEXT:    br i1 [[TMP3]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, i32* [[CONV]], align 8
// CHECK2-NEXT:    [[TMP5:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 1, i64 80, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.1*)* @.omp_task_entry..4 to i32 (i32, i8*)*))
// CHECK2-NEXT:    [[TMP6:%.*]] = bitcast i8* [[TMP5]] to %struct.kmp_task_t_with_privates.1*
// CHECK2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_1:%.*]], %struct.kmp_task_t_with_privates.1* [[TMP6]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 5
// CHECK2-NEXT:    store i64 0, i64* [[TMP8]], align 8
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 6
// CHECK2-NEXT:    store i64 9, i64* [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 7
// CHECK2-NEXT:    store i64 1, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP7]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP12:%.*]] = bitcast i8** [[TMP11]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP12]], i8 0, i64 8, i1 false)
// CHECK2-NEXT:    [[TMP13:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP14:%.*]] = zext i32 [[TMP4]] to i64
// CHECK2-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i8* [[TMP5]], i32 1, i64* [[TMP8]], i64* [[TMP9]], i64 [[TMP13]], i32 1, i32 1, i64 [[TMP14]], i8* null)
// CHECK2-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_entry..4
// CHECK2-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.1* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.0*, 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.1*, align 8
// CHECK2-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    store %struct.kmp_task_t_with_privates.1* [[TMP1]], %struct.kmp_task_t_with_privates.1** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.1*, %struct.kmp_task_t_with_privates.1** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_1:%.*]], %struct.kmp_task_t_with_privates.1* [[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.0*
// CHECK2-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.1* [[TMP3]] to i8*
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK2-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META17:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META20:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META22:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META24:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META26:![0-9]+]])
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !28
// CHECK2-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !28
// CHECK2-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    store %struct.anon.0* [[TMP8]], %struct.anon.0** [[__CONTEXT_ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    [[TMP20:%.*]] = load %struct.anon.0*, %struct.anon.0** [[__CONTEXT_ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    [[TMP21:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP21]] to i32
// CHECK2-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK2:       omp.inner.for.cond.i:
// CHECK2-NEXT:    [[TMP22:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK2-NEXT:    [[CONV1_I:%.*]] = sext i32 [[TMP22]] to i64
// CHECK2-NEXT:    [[TMP23:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !28
// CHECK2-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV1_I]], [[TMP23]]
// CHECK2-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__3_EXIT:%.*]]
// CHECK2:       omp.inner.for.body.i:
// CHECK2-NEXT:    [[TMP24:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK2-NEXT:    store i32 [[TMP24]], i32* [[I_I]], align 4, !noalias !28
// CHECK2-NEXT:    [[TMP25:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK2-NEXT:    [[ADD2_I:%.*]] = add nsw i32 [[TMP25]], 1
// CHECK2-NEXT:    store i32 [[ADD2_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !28
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK2:       .omp_outlined..3.exit:
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined..5
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i8*** nonnull align 8 dereferenceable(8) [[ARGV:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]], i64 [[DOTCAPTURE_EXPR_1:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[ARGV_ADDR:%.*]] = alloca i8***, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR2:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_2:%.*]], align 8
// CHECK2-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[_TMP4:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_5:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_6:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_7:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_11:%.*]] = alloca i64, align 8
// 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* [[ARGC]], i32** [[ARGC_ADDR]], align 8
// CHECK2-NEXT:    store i8*** [[ARGV]], i8**** [[ARGV_ADDR]], align 8
// CHECK2-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK2-NEXT:    store i64 [[DOTCAPTURE_EXPR_1]], i64* [[DOTCAPTURE_EXPR__ADDR2]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[ARGC_ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i8***, i8**** [[ARGV_ADDR]], align 8
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i8*
// CHECK2-NEXT:    [[CONV3:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR2]] to i32*
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    [[TMP5:%.*]] = icmp ne i32 [[TMP4]], 0
// CHECK2-NEXT:    br i1 [[TMP5]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[AGG_CAPTURED]], i32 0, i32 0
// CHECK2-NEXT:    store i32* [[TMP0]], i32** [[TMP6]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[AGG_CAPTURED]], i32 0, i32 1
// CHECK2-NEXT:    store i8*** [[TMP1]], i8**** [[TMP7]], align 8
// CHECK2-NEXT:    [[TMP8:%.*]] = load i32, i32* [[CONV3]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    store i32 [[TMP9]], i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK2-NEXT:    [[TMP10:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    store i32 [[TMP10]], i32* [[DOTCAPTURE_EXPR_6]], align 4
// CHECK2-NEXT:    [[TMP11:%.*]] = load i8**, i8*** [[TMP1]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP12]] to i64
// CHECK2-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i8*, i8** [[TMP11]], i64 [[IDXPROM]]
// CHECK2-NEXT:    [[TMP13:%.*]] = load i8*, i8** [[ARRAYIDX]], align 8
// CHECK2-NEXT:    [[TMP14:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[TMP14]] to i64
// CHECK2-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds i8, i8* [[TMP13]], i64 [[IDXPROM8]]
// CHECK2-NEXT:    [[TMP15:%.*]] = load i8, i8* [[ARRAYIDX9]], align 1
// CHECK2-NEXT:    [[CONV10:%.*]] = sext i8 [[TMP15]] to i32
// CHECK2-NEXT:    store i32 [[CONV10]], i32* [[DOTCAPTURE_EXPR_7]], align 4
// CHECK2-NEXT:    [[TMP16:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_5]], align 4
// CHECK2-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP16]], 0
// CHECK2-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK2-NEXT:    [[CONV12:%.*]] = sext i32 [[DIV]] to i64
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_7]], align 4
// CHECK2-NEXT:    [[TMP18:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_6]], align 4
// CHECK2-NEXT:    [[SUB13:%.*]] = sub i32 [[TMP17]], [[TMP18]]
// CHECK2-NEXT:    [[SUB14:%.*]] = sub i32 [[SUB13]], 1
// CHECK2-NEXT:    [[ADD:%.*]] = add i32 [[SUB14]], 1
// CHECK2-NEXT:    [[DIV15:%.*]] = udiv i32 [[ADD]], 1
// CHECK2-NEXT:    [[CONV16:%.*]] = zext i32 [[DIV15]] to i64
// CHECK2-NEXT:    [[MUL:%.*]] = mul nsw i64 [[CONV12]], [[CONV16]]
// CHECK2-NEXT:    [[SUB17:%.*]] = sub nsw i64 [[MUL]], 1
// CHECK2-NEXT:    store i64 [[SUB17]], i64* [[DOTCAPTURE_EXPR_11]], align 8
// CHECK2-NEXT:    [[TMP19:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i32 1, i64 80, i64 16, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.3*)* @.omp_task_entry..7 to i32 (i32, i8*)*))
// CHECK2-NEXT:    [[TMP20:%.*]] = bitcast i8* [[TMP19]] to %struct.kmp_task_t_with_privates.3*
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_3:%.*]], %struct.kmp_task_t_with_privates.3* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP23:%.*]] = load i8*, i8** [[TMP22]], align 8
// CHECK2-NEXT:    [[TMP24:%.*]] = bitcast %struct.anon.2* [[AGG_CAPTURED]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP23]], i8* align 8 [[TMP24]], i64 16, i1 false)
// CHECK2-NEXT:    [[TMP25:%.*]] = load i8, i8* [[CONV]], align 8
// CHECK2-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[TMP25]] to i1
// CHECK2-NEXT:    [[TMP26:%.*]] = sext i1 [[TOBOOL]] to i32
// CHECK2-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 5
// CHECK2-NEXT:    store i64 0, i64* [[TMP27]], align 8
// CHECK2-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP29:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR_11]], align 8
// CHECK2-NEXT:    store i64 [[TMP29]], i64* [[TMP28]], align 8
// CHECK2-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 7
// CHECK2-NEXT:    store i64 1, i64* [[TMP30]], align 8
// CHECK2-NEXT:    [[TMP31:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP21]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP32:%.*]] = bitcast i8** [[TMP31]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP32]], i8 0, i64 8, i1 false)
// CHECK2-NEXT:    [[TMP33:%.*]] = load i64, i64* [[TMP30]], align 8
// CHECK2-NEXT:    [[TMP34:%.*]] = zext i32 [[TMP8]] to i64
// CHECK2-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i8* [[TMP19]], i32 [[TMP26]], i64* [[TMP27]], i64* [[TMP28]], i64 [[TMP33]], i32 1, i32 2, i64 [[TMP34]], i8* null)
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_entry..7
// CHECK2-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.3* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.2*, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_2_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_3_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_6_I:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[J_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[I14_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[J15_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca %struct.kmp_task_t_with_privates.3*, align 8
// CHECK2-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    store %struct.kmp_task_t_with_privates.3* [[TMP1]], %struct.kmp_task_t_with_privates.3** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.3*, %struct.kmp_task_t_with_privates.3** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_3:%.*]], %struct.kmp_task_t_with_privates.3* [[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.2*
// CHECK2-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.3* [[TMP3]] to i8*
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK2-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META29:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META32:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META34:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META36:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META38:![0-9]+]])
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !40
// CHECK2-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !40
// CHECK2-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store %struct.anon.2* [[TMP8]], %struct.anon.2** [[__CONTEXT_ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP20:%.*]] = load %struct.anon.2*, %struct.anon.2** [[__CONTEXT_ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_ANON_2:%.*]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP22:%.*]] = load i32*, i32** [[TMP21]], align 8
// CHECK2-NEXT:    [[TMP23:%.*]] = load i32, i32* [[TMP22]], align 4
// CHECK2-NEXT:    store i32 [[TMP23]], i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[TMP24]], align 8
// CHECK2-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
// CHECK2-NEXT:    store i32 [[TMP26]], i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP28:%.*]] = load i8***, i8**** [[TMP27]], align 8
// CHECK2-NEXT:    [[TMP29:%.*]] = load i8**, i8*** [[TMP28]], align 8
// CHECK2-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP31:%.*]] = load i32*, i32** [[TMP30]], align 8
// CHECK2-NEXT:    [[TMP32:%.*]] = load i32, i32* [[TMP31]], align 4
// CHECK2-NEXT:    [[IDXPROM_I:%.*]] = sext i32 [[TMP32]] to i64
// CHECK2-NEXT:    [[ARRAYIDX_I:%.*]] = getelementptr inbounds i8*, i8** [[TMP29]], i64 [[IDXPROM_I]]
// CHECK2-NEXT:    [[TMP33:%.*]] = load i8*, i8** [[ARRAYIDX_I]], align 8
// CHECK2-NEXT:    [[TMP34:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP35:%.*]] = load i32*, i32** [[TMP34]], align 8
// CHECK2-NEXT:    [[TMP36:%.*]] = load i32, i32* [[TMP35]], align 4
// CHECK2-NEXT:    [[IDXPROM4_I:%.*]] = sext i32 [[TMP36]] to i64
// CHECK2-NEXT:    [[ARRAYIDX5_I:%.*]] = getelementptr inbounds i8, i8* [[TMP33]], i64 [[IDXPROM4_I]]
// CHECK2-NEXT:    [[TMP37:%.*]] = load i8, i8* [[ARRAYIDX5_I]], align 1
// CHECK2-NEXT:    [[CONV_I:%.*]] = sext i8 [[TMP37]] to i32
// CHECK2-NEXT:    store i32 [[CONV_I]], i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP38:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !40
// CHECK2-NEXT:    [[CONV7_I:%.*]] = sext i32 [[TMP38]] to i64
// CHECK2-NEXT:    [[TMP39:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP40:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[SUB8_I:%.*]] = sub i32 [[TMP39]], [[TMP40]]
// CHECK2-NEXT:    [[SUB9_I:%.*]] = sub i32 [[SUB8_I]], 1
// CHECK2-NEXT:    [[CONV11_I:%.*]] = zext i32 [[SUB8_I]] to i64
// CHECK2-NEXT:    [[MUL_I:%.*]] = mul nsw i64 [[CONV7_I]], [[CONV11_I]]
// CHECK2-NEXT:    [[SUB12_I:%.*]] = sub nsw i64 [[MUL_I]], 1
// CHECK2-NEXT:    store i64 [[SUB12_I]], i64* [[DOTCAPTURE_EXPR_6_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i32 0, i32* [[I_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP41:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    store i32 [[TMP41]], i32* [[J_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP42:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !40
// CHECK2-NEXT:    [[CMP_I:%.*]] = icmp slt i32 0, [[TMP42]]
// CHECK2-NEXT:    br i1 [[CMP_I]], label [[LAND_LHS_TRUE_I:%.*]], label [[DOTOMP_OUTLINED__6_EXIT:%.*]]
// CHECK2:       land.lhs.true.i:
// CHECK2-NEXT:    [[TMP43:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP44:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[CMP13_I:%.*]] = icmp slt i32 [[TMP43]], [[TMP44]]
// CHECK2-NEXT:    br i1 [[CMP13_I]], label [[TASKLOOP_IF_THEN_I:%.*]], label [[DOTOMP_OUTLINED__6_EXIT]]
// CHECK2:       taskloop.if.then.i:
// CHECK2-NEXT:    [[TMP45:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    store i64 [[TMP45]], i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP46:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP47:%.*]] = load i32*, i32** [[TMP46]], align 8
// CHECK2-NEXT:    [[TMP48:%.*]] = getelementptr inbounds [[STRUCT_ANON_2]], %struct.anon.2* [[TMP20]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP49:%.*]] = load i8***, i8**** [[TMP48]], align 8
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK2:       omp.inner.for.cond.i:
// CHECK2-NEXT:    [[TMP50:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP51:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[CMP16_I:%.*]] = icmp ule i64 [[TMP50]], [[TMP51]]
// CHECK2-NEXT:    br i1 [[CMP16_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[OMP_INNER_FOR_END_I:%.*]]
// CHECK2:       omp.inner.for.body.i:
// CHECK2-NEXT:    [[TMP52:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP53:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP54:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[SUB17_I:%.*]] = sub i32 [[TMP53]], [[TMP54]]
// CHECK2-NEXT:    [[SUB18_I:%.*]] = sub i32 [[SUB17_I]], 1
// CHECK2-NEXT:    [[CONV22_I:%.*]] = zext i32 [[SUB17_I]] to i64
// CHECK2-NEXT:    [[DIV23_I:%.*]] = sdiv i64 [[TMP52]], [[CONV22_I]]
// CHECK2-NEXT:    [[CONV26_I:%.*]] = trunc i64 [[DIV23_I]] to i32
// CHECK2-NEXT:    store i32 [[CONV26_I]], i32* [[I14_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP55:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[CONV27_I:%.*]] = sext i32 [[TMP55]] to i64
// CHECK2-NEXT:    [[TMP56:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP57:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[TMP58:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP59:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[SUB28_I:%.*]] = sub i32 [[TMP58]], [[TMP59]]
// CHECK2-NEXT:    [[SUB29_I:%.*]] = sub i32 [[SUB28_I]], 1
// CHECK2-NEXT:    [[CONV33_I:%.*]] = zext i32 [[SUB28_I]] to i64
// CHECK2-NEXT:    [[DIV34_I:%.*]] = sdiv i64 [[TMP57]], [[CONV33_I]]
// CHECK2-NEXT:    [[TMP60:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP61:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[SUB35_I:%.*]] = sub i32 [[TMP60]], [[TMP61]]
// CHECK2-NEXT:    [[SUB36_I:%.*]] = sub i32 [[SUB35_I]], 1
// CHECK2-NEXT:    [[CONV40_I:%.*]] = zext i32 [[SUB35_I]] to i64
// CHECK2-NEXT:    [[MUL41_I:%.*]] = mul nsw i64 [[DIV34_I]], [[CONV40_I]]
// CHECK2-NEXT:    [[SUB42_I:%.*]] = sub nsw i64 [[TMP56]], [[MUL41_I]]
// CHECK2-NEXT:    [[ADD44_I:%.*]] = add nsw i64 [[CONV27_I]], [[SUB42_I]]
// CHECK2-NEXT:    [[CONV45_I:%.*]] = trunc i64 [[ADD44_I]] to i32
// CHECK2-NEXT:    store i32 [[CONV45_I]], i32* [[J15_I]], align 4, !noalias !40
// CHECK2-NEXT:    [[TMP62:%.*]] = load i64, i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    [[ADD46_I:%.*]] = add nsw i64 [[TMP62]], 1
// CHECK2-NEXT:    store i64 [[ADD46_I]], i64* [[DOTOMP_IV_I]], align 8, !noalias !40
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK2:       omp.inner.for.end.i:
// CHECK2-NEXT:    br label [[DOTOMP_OUTLINED__6_EXIT]]
// CHECK2:       .omp_outlined..6.exit:
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined..8
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_4:%.*]], align 1
// 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:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    [[TMP3:%.*]] = icmp ne i32 [[TMP2]], 0
// CHECK2-NEXT:    br i1 [[TMP3]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    [[TMP4:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 1, i64 80, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.5*)* @.omp_task_entry..10 to i32 (i32, i8*)*))
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i8* [[TMP4]] to %struct.kmp_task_t_with_privates.5*
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_5:%.*]], %struct.kmp_task_t_with_privates.5* [[TMP5]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 5
// CHECK2-NEXT:    store i64 0, i64* [[TMP7]], align 8
// CHECK2-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 6
// CHECK2-NEXT:    store i64 9, i64* [[TMP8]], align 8
// CHECK2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 7
// CHECK2-NEXT:    store i64 1, i64* [[TMP9]], align 8
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP6]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP11:%.*]] = bitcast i8** [[TMP10]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP11]], i8 0, i64 8, i1 false)
// CHECK2-NEXT:    [[TMP12:%.*]] = load i64, i64* [[TMP9]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i8* [[TMP4]], i32 1, i64* [[TMP7]], i64* [[TMP8]], i64 [[TMP12]], i32 1, i32 0, i64 0, i8* null)
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_entry..10
// CHECK2-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.5* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.4*, align 8
// CHECK2-NEXT:    [[I_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IV_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[CLEANUP_DEST_SLOT_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca %struct.kmp_task_t_with_privates.5*, align 8
// CHECK2-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    store %struct.kmp_task_t_with_privates.5* [[TMP1]], %struct.kmp_task_t_with_privates.5** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.5*, %struct.kmp_task_t_with_privates.5** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_5:%.*]], %struct.kmp_task_t_with_privates.5* [[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.4*
// CHECK2-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.5* [[TMP3]] to i8*
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK2-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META41:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META44:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META46:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META48:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META50:![0-9]+]])
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !52
// CHECK2-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !52
// CHECK2-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    store %struct.anon.4* [[TMP8]], %struct.anon.4** [[__CONTEXT_ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    [[TMP20:%.*]] = load %struct.anon.4*, %struct.anon.4** [[__CONTEXT_ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    [[TMP21:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP21]] to i32
// CHECK2-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK2:       omp.inner.for.cond.i:
// CHECK2-NEXT:    [[TMP22:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK2-NEXT:    [[CONV1_I:%.*]] = sext i32 [[TMP22]] to i64
// CHECK2-NEXT:    [[TMP23:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !52
// CHECK2-NEXT:    [[CMP_I:%.*]] = icmp ule i64 [[CONV1_I]], [[TMP23]]
// CHECK2-NEXT:    br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[OMP_INNER_FOR_END_I:%.*]]
// CHECK2:       omp.inner.for.body.i:
// CHECK2-NEXT:    [[TMP24:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK2-NEXT:    store i32 [[TMP24]], i32* [[I_I]], align 4, !noalias !52
// CHECK2-NEXT:    [[TMP25:%.*]] = load i32, i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !52
// CHECK2-NEXT:    [[TMP26:%.*]] = call i32 @__kmpc_cancel(%struct.ident_t* @[[GLOB1]], i32 [[TMP25]], i32 4) #[[ATTR2]]
// CHECK2-NEXT:    [[TMP27:%.*]] = icmp ne i32 [[TMP26]], 0
// CHECK2-NEXT:    br i1 [[TMP27]], label [[DOTCANCEL_EXIT_I:%.*]], label [[DOTCANCEL_CONTINUE_I:%.*]]
// CHECK2:       .cancel.exit.i:
// CHECK2-NEXT:    store i32 1, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK2-NEXT:    br label [[DOTOMP_OUTLINED__9_EXIT:%.*]]
// CHECK2:       .cancel.continue.i:
// CHECK2-NEXT:    [[TMP28:%.*]] = load i32, i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !52
// CHECK2-NEXT:    [[TMP29:%.*]] = call i32 @__kmpc_cancellationpoint(%struct.ident_t* @[[GLOB1]], i32 [[TMP28]], i32 4) #[[ATTR2]]
// CHECK2-NEXT:    [[TMP30:%.*]] = icmp ne i32 [[TMP29]], 0
// CHECK2-NEXT:    br i1 [[TMP30]], label [[DOTCANCEL_EXIT2_I:%.*]], label [[DOTCANCEL_CONTINUE3_I:%.*]]
// CHECK2:       .cancel.exit2.i:
// CHECK2-NEXT:    store i32 1, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK2-NEXT:    br label [[DOTOMP_OUTLINED__9_EXIT]]
// CHECK2:       .cancel.continue3.i:
// CHECK2-NEXT:    [[TMP31:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK2-NEXT:    [[ADD4_I:%.*]] = add nsw i32 [[TMP31]], 1
// CHECK2-NEXT:    store i32 [[ADD4_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !52
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK2:       omp.inner.for.end.i:
// CHECK2-NEXT:    store i32 0, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK2-NEXT:    br label [[DOTOMP_OUTLINED__9_EXIT]]
// CHECK2:       .omp_outlined..9.exit:
// CHECK2-NEXT:    [[CLEANUP_DEST_I:%.*]] = load i32, i32* [[CLEANUP_DEST_SLOT_I]], align 4, !noalias !52
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@__cxx_global_var_init
// CHECK2-SAME: () #[[ATTR7:[0-9]+]] section "__TEXT,__StaticInit,regular,pure_instructions" {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    call void @_ZN1SC1Ei(%struct.S* nonnull align 4 dereferenceable(4) @s, i32 1)
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SC1Ei
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[C:%.*]]) unnamed_addr #[[ATTR8:[0-9]+]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[C_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store i32 [[C]], i32* [[C_ADDR]], align 4
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32, i32* [[C_ADDR]], align 4
// CHECK2-NEXT:    call void @_ZN1SC2Ei(%struct.S* nonnull align 4 dereferenceable(4) [[THIS1]], i32 [[TMP0]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_ZN1SC2Ei
// CHECK2-SAME: (%struct.S* nonnull align 4 dereferenceable(4) [[THIS:%.*]], i32 [[C:%.*]]) unnamed_addr #[[ATTR8]] align 2 {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[C_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i8, align 1
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    store %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store i32 [[C]], i32* [[C_ADDR]], align 4
// CHECK2-NEXT:    [[THIS1:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32, i32* [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TOBOOL:%.*]] = icmp ne i32 [[TMP0]], 0
// CHECK2-NEXT:    [[FROMBOOL:%.*]] = zext i1 [[TOBOOL]] to i8
// CHECK2-NEXT:    store i8 [[FROMBOOL]], i8* [[DOTCAPTURE_EXPR_]], align 1
// CHECK2-NEXT:    [[TMP1:%.*]] = load i8, i8* [[DOTCAPTURE_EXPR_]], align 1
// CHECK2-NEXT:    [[TOBOOL2:%.*]] = trunc i8 [[TMP1]] to i1
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__CASTED]] to i8*
// CHECK2-NEXT:    [[FROMBOOL3:%.*]] = zext i1 [[TOBOOL2]] to i8
// CHECK2-NEXT:    store i8 [[FROMBOOL3]], i8* [[CONV]], align 1
// CHECK2-NEXT:    [[TMP2:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK2-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, %struct.S*, i32*, i64)* @.omp_outlined..11 to void (i32*, i32*, ...)*), %struct.S* [[THIS1]], i32* [[C_ADDR]], i64 [[TMP2]])
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_outlined..11
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], %struct.S* [[THIS:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK2-NEXT:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON_6:%.*]], align 8
// CHECK2-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[_TMP1:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_3:%.*]] = 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 %struct.S* [[THIS]], %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 8
// CHECK2-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], i64* [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK2-NEXT:    [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[C_ADDR]], align 8
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i8*
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = call i32 @__kmpc_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    [[TMP5:%.*]] = icmp ne i32 [[TMP4]], 0
// CHECK2-NEXT:    br i1 [[TMP5]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK2:       omp_if.then:
// CHECK2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[AGG_CAPTURED]], i32 0, i32 0
// CHECK2-NEXT:    store %struct.S* [[TMP0]], %struct.S** [[TMP6]], align 8
// CHECK2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[AGG_CAPTURED]], i32 0, i32 1
// CHECK2-NEXT:    store i32* [[TMP1]], i32** [[TMP7]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    [[TMP8:%.*]] = load i8, i8* [[CONV]], align 8
// CHECK2-NEXT:    [[TOBOOL:%.*]] = trunc i8 [[TMP8]] to i1
// CHECK2-NEXT:    store i32* [[TMP]], i32** [[_TMP1]], align 8
// CHECK2-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP1]], align 4
// CHECK2-NEXT:    store i32 [[TMP9]], i32* [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    [[TMP10:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP10]], 0
// CHECK2-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK2-NEXT:    [[SUB4:%.*]] = sub nsw i32 [[DIV]], 1
// CHECK2-NEXT:    store i32 [[SUB4]], i32* [[DOTCAPTURE_EXPR_3]], align 4
// CHECK2-NEXT:    [[TMP11:%.*]] = select i1 [[TOBOOL]], i32 2, i32 0
// CHECK2-NEXT:    [[TMP12:%.*]] = or i32 [[TMP11]], 1
// CHECK2-NEXT:    [[TMP13:%.*]] = call i8* @__kmpc_omp_task_alloc(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i32 [[TMP12]], i64 80, i64 16, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates.7*)* @.omp_task_entry..13 to i32 (i32, i8*)*))
// CHECK2-NEXT:    [[TMP14:%.*]] = bitcast i8* [[TMP13]] to %struct.kmp_task_t_with_privates.7*
// CHECK2-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_7:%.*]], %struct.kmp_task_t_with_privates.7* [[TMP14]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T:%.*]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP17:%.*]] = load i8*, i8** [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = bitcast %struct.anon.6* [[AGG_CAPTURED]] to i8*
// CHECK2-NEXT:    call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 8 [[TMP17]], i8* align 8 [[TMP18]], i64 16, i1 false)
// CHECK2-NEXT:    [[TMP19:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 5
// CHECK2-NEXT:    store i64 0, i64* [[TMP19]], align 8
// CHECK2-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 6
// CHECK2-NEXT:    [[TMP21:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3]], align 4
// CHECK2-NEXT:    [[CONV5:%.*]] = sext i32 [[TMP21]] to i64
// CHECK2-NEXT:    store i64 [[CONV5]], i64* [[TMP20]], align 8
// CHECK2-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 7
// CHECK2-NEXT:    store i64 1, i64* [[TMP22]], align 8
// CHECK2-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP15]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP24:%.*]] = bitcast i8** [[TMP23]] to i8*
// CHECK2-NEXT:    call void @llvm.memset.p0i8.i64(i8* align 8 [[TMP24]], i8 0, i64 8, i1 false)
// CHECK2-NEXT:    [[TMP25:%.*]] = load i64, i64* [[TMP22]], align 8
// CHECK2-NEXT:    call void @__kmpc_taskloop(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]], i8* [[TMP13]], i32 1, i64* [[TMP19]], i64* [[TMP20]], i64 [[TMP25]], i32 1, i32 2, i64 4, i8* null)
// CHECK2-NEXT:    call void @__kmpc_end_taskgroup(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    call void @__kmpc_end_master(%struct.ident_t* @[[GLOB1]], i32 [[TMP3]])
// CHECK2-NEXT:    br label [[OMP_IF_END]]
// CHECK2:       omp_if.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@.omp_task_entry..13
// CHECK2-SAME: (i32 [[TMP0:%.*]], %struct.kmp_task_t_with_privates.7* noalias [[TMP1:%.*]]) #[[ATTR4]] {
// 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.6*, align 8
// CHECK2-NEXT:    [[TMP_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TMP1_I:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_2_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[A_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TMP4_I:%.*]] = alloca i32*, align 8
// CHECK2-NEXT:    [[A5_I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TMP6_I:%.*]] = alloca i32*, align 8
// 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.7*, align 8
// CHECK2-NEXT:    store i32 [[TMP0]], i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    store %struct.kmp_task_t_with_privates.7* [[TMP1]], %struct.kmp_task_t_with_privates.7** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTADDR]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load %struct.kmp_task_t_with_privates.7*, %struct.kmp_task_t_with_privates.7** [[DOTADDR1]], align 8
// CHECK2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T_WITH_PRIVATES_7:%.*]], %struct.kmp_task_t_with_privates.7* [[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.6*
// CHECK2-NEXT:    [[TMP9:%.*]] = bitcast %struct.kmp_task_t_with_privates.7* [[TMP3]] to i8*
// CHECK2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 5
// CHECK2-NEXT:    [[TMP11:%.*]] = load i64, i64* [[TMP10]], align 8
// CHECK2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 6
// 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 7
// 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 8
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 8
// CHECK2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT_KMP_TASK_T]], %struct.kmp_task_t* [[TMP4]], i32 0, i32 9
// CHECK2-NEXT:    [[TMP19:%.*]] = load i8*, i8** [[TMP18]], align 8
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META53:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META56:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META58:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META60:![0-9]+]])
// CHECK2-NEXT:    call void @llvm.experimental.noalias.scope.decl(metadata [[META62:![0-9]+]])
// CHECK2-NEXT:    store i32 [[TMP2]], i32* [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias !64
// CHECK2-NEXT:    store i32* [[TMP5]], i32** [[DOTPART_ID__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i8* null, i8** [[DOTPRIVATES__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store void (i8*, ...)* null, void (i8*, ...)** [[DOTCOPY_FN__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i8* [[TMP9]], i8** [[DOTTASK_T__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i64 [[TMP11]], i64* [[DOTLB__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i64 [[TMP13]], i64* [[DOTUB__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i64 [[TMP15]], i64* [[DOTST__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i32 [[TMP17]], i32* [[DOTLITER__ADDR_I]], align 4, !noalias !64
// CHECK2-NEXT:    store i8* [[TMP19]], i8** [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    store %struct.anon.6* [[TMP8]], %struct.anon.6** [[__CONTEXT_ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[TMP20:%.*]] = load %struct.anon.6*, %struct.anon.6** [[__CONTEXT_ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [[STRUCT_ANON_6:%.*]], %struct.anon.6* [[TMP20]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP22:%.*]] = load %struct.S*, %struct.S** [[TMP21]], align 8
// CHECK2-NEXT:    store i32* [[TMP_I]], i32** [[TMP1_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[TMP20]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP24:%.*]] = load i32*, i32** [[TMP23]], align 8
// CHECK2-NEXT:    [[TMP25:%.*]] = load i32, i32* [[TMP24]], align 4
// CHECK2-NEXT:    store i32 [[TMP25]], i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !64
// CHECK2-NEXT:    [[TMP26:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !64
// CHECK2-NEXT:    [[SUB3_I:%.*]] = sub nsw i32 [[TMP26]], 1
// CHECK2-NEXT:    store i32 [[SUB3_I]], i32* [[DOTCAPTURE_EXPR_2_I]], align 4, !noalias !64
// CHECK2-NEXT:    store i32* [[A_I]], i32** [[TMP4_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[TMP27:%.*]] = load i32*, i32** [[TMP4_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i32 0, i32* [[TMP27]], align 4
// CHECK2-NEXT:    [[TMP28:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__I]], align 4, !noalias !64
// CHECK2-NEXT:    [[CMP_I:%.*]] = icmp slt i32 0, [[TMP28]]
// CHECK2-NEXT:    br i1 [[CMP_I]], label [[TASKLOOP_IF_THEN_I:%.*]], label [[DOTOMP_OUTLINED__12_EXIT:%.*]]
// CHECK2:       taskloop.if.then.i:
// CHECK2-NEXT:    store i32* [[A5_I]], i32** [[TMP6_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[TMP29:%.*]] = load i64, i64* [[DOTLB__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[CONV_I:%.*]] = trunc i64 [[TMP29]] to i32
// CHECK2-NEXT:    store i32 [[CONV_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK2-NEXT:    [[TMP30:%.*]] = getelementptr inbounds [[STRUCT_ANON_6]], %struct.anon.6* [[TMP20]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP31:%.*]] = load i32*, i32** [[TMP30]], align 8
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK2:       omp.inner.for.cond.i:
// CHECK2-NEXT:    [[TMP32:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK2-NEXT:    [[CONV7_I:%.*]] = sext i32 [[TMP32]] to i64
// CHECK2-NEXT:    [[TMP33:%.*]] = load i64, i64* [[DOTUB__ADDR_I]], align 8, !noalias !64
// CHECK2-NEXT:    [[CMP8_I:%.*]] = icmp ule i64 [[CONV7_I]], [[TMP33]]
// CHECK2-NEXT:    br i1 [[CMP8_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[OMP_INNER_FOR_END_I:%.*]]
// CHECK2:       omp.inner.for.body.i:
// CHECK2-NEXT:    [[TMP34:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK2-NEXT:    [[TMP35:%.*]] = load i32*, i32** [[TMP6_I]], align 8, !noalias !64
// CHECK2-NEXT:    store i32 [[TMP34]], i32* [[TMP35]], align 4
// CHECK2-NEXT:    [[TMP36:%.*]] = load i32, i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK2-NEXT:    [[ADD9_I:%.*]] = add nsw i32 [[TMP36]], 1
// CHECK2-NEXT:    store i32 [[ADD9_I]], i32* [[DOTOMP_IV_I]], align 4, !noalias !64
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND_I]]
// CHECK2:       omp.inner.for.end.i:
// CHECK2-NEXT:    br label [[DOTOMP_OUTLINED__12_EXIT]]
// CHECK2:       .omp_outlined..12.exit:
// CHECK2-NEXT:    ret i32 0
//
//
// CHECK2-LABEL: define {{[^@]+}}@_GLOBAL__sub_I_parallel_master_taskloop_codegen.cpp
// CHECK2-SAME: () #[[ATTR7]] section "__TEXT,__StaticInit,regular,pure_instructions" {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    call void @__cxx_global_var_init()
// CHECK2-NEXT:    ret void
//