// 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 _
// Test target codegen - host bc file has to be created first.
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix=CHECK2
// RUN: %clang_cc1 -verify -fopenmp -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix=CHECK3
// expected-no-diagnostics
#ifndef HEADER
#define HEADER

template<typename tx>
tx ftemplate(int n) {
  tx a = 0;
  short aa = 0;
  tx b[10];

  #pragma omp target teams if(0)
  {
    b[2] += 1;
  }

  #pragma omp target teams if(1)
  {
    a = '1';
  }

  #pragma omp target teams if(n>40)
  {
    aa = 1;
  }

  #pragma omp target teams
  {
#pragma omp parallel
#pragma omp parallel
    aa = 1;
  }

  return a;
}

int bar(int n){
  int a = 0;

  a += ftemplate<char>(n);

  return a;
}

#endif
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l23
// CHECK1-SAME: (i64 [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[A_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT:    store i64 [[A]], i64* [[A_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i8*
// CHECK1-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i1 false, i1 true, i1 true)
// CHECK1-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK1-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK1:       user_code.entry:
// CHECK1-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT:    [[TMP2:%.*]] = load i8, i8* [[CONV]], align 8
// CHECK1-NEXT:    [[CONV1:%.*]] = bitcast i64* [[A_CASTED]] to i8*
// CHECK1-NEXT:    store i8 [[TMP2]], i8* [[CONV1]], align 1
// CHECK1-NEXT:    [[TMP3:%.*]] = load i64, i64* [[A_CASTED]], align 8
// CHECK1-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT:    call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP3]]) #[[ATTR1:[0-9]+]]
// CHECK1-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
// CHECK1-NEXT:    ret void
// CHECK1:       worker.exit:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[A:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca 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 i64 [[A]], i64* [[A_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i8*
// CHECK1-NEXT:    store i8 49, i8* [[CONV]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l28
// CHECK1-SAME: (i64 [[AA:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[AA_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[AA_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT:    store i64 [[AA]], i64* [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16*
// CHECK1-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 false, i1 true, i1 true)
// CHECK1-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK1-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK1:       user_code.entry:
// CHECK1-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT:    [[TMP2:%.*]] = load i16, i16* [[CONV]], align 8
// CHECK1-NEXT:    [[CONV1:%.*]] = bitcast i64* [[AA_CASTED]] to i16*
// CHECK1-NEXT:    store i16 [[TMP2]], i16* [[CONV1]], align 2
// CHECK1-NEXT:    [[TMP3:%.*]] = load i64, i64* [[AA_CASTED]], align 8
// CHECK1-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT:    call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP3]]) #[[ATTR1]]
// CHECK1-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
// CHECK1-NEXT:    ret void
// CHECK1:       worker.exit:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[AA:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[AA_ADDR:%.*]] = 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 i64 [[AA]], i64* [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16*
// CHECK1-NEXT:    store i16 1, i16* [[CONV]], align 8
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33
// CHECK1-SAME: (i64 [[AA:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[AA_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[AA_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT:    store i64 [[AA]], i64* [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16*
// CHECK1-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 true)
// CHECK1-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK1-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK1:       user_code.entry:
// CHECK1-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK1-NEXT:    [[TMP2:%.*]] = load i16, i16* [[CONV]], align 8
// CHECK1-NEXT:    [[CONV1:%.*]] = bitcast i64* [[AA_CASTED]] to i16*
// CHECK1-NEXT:    store i16 [[TMP2]], i16* [[CONV1]], align 2
// CHECK1-NEXT:    [[TMP3:%.*]] = load i64, i64* [[AA_CASTED]], align 8
// CHECK1-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT:    call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP3]]) #[[ATTR1]]
// CHECK1-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
// CHECK1-NEXT:    ret void
// CHECK1:       worker.exit:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__2
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i64 [[AA:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[AA_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], 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 i64 [[AA]], i64* [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16*
// CHECK1-NEXT:    [[TMP0:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP1:%.*]] = bitcast i16* [[CONV]] to i8*
// CHECK1-NEXT:    store i8* [[TMP1]], i8** [[TMP0]], align 8
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP3]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__3 to i8*), i8* null, i8** [[TMP4]], i64 1)
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__3
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i16* nonnull align 2 dereferenceable(2) [[AA:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[AA_ADDR:%.*]] = alloca i16*, align 8
// CHECK1-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], 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 i16* [[AA]], i16** [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP2:%.*]] = bitcast i16* [[TMP0]] to i8*
// CHECK1-NEXT:    store i8* [[TMP2]], i8** [[TMP1]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__4 to i8*), i8* null, i8** [[TMP5]], i64 1)
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__4
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i16* nonnull align 2 dereferenceable(2) [[AA:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[AA_ADDR:%.*]] = alloca i16*, 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 i16* [[AA]], i16** [[AA_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 8
// CHECK1-NEXT:    store i16 1, i16* [[TMP0]], align 2
// CHECK1-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l23
// CHECK2-SAME: (i32 [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[A_CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i32* [[A_ADDR]] to i8*
// CHECK2-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i1 false, i1 true, i1 true)
// CHECK2-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK2-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK2:       user_code.entry:
// CHECK2-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT:    [[TMP2:%.*]] = load i8, i8* [[CONV]], align 4
// CHECK2-NEXT:    [[CONV1:%.*]] = bitcast i32* [[A_CASTED]] to i8*
// CHECK2-NEXT:    store i8 [[TMP2]], i8* [[CONV1]], align 1
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[A_CASTED]], align 4
// CHECK2-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT:    call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1:[0-9]+]]
// CHECK2-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
// CHECK2-NEXT:    ret void
// CHECK2:       worker.exit:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32 [[A:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i32* [[A_ADDR]] to i8*
// CHECK2-NEXT:    store i8 49, i8* [[CONV]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l28
// CHECK2-SAME: (i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[AA_CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK2-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 false, i1 true, i1 true)
// CHECK2-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK2-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK2:       user_code.entry:
// CHECK2-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT:    [[TMP2:%.*]] = load i16, i16* [[CONV]], align 4
// CHECK2-NEXT:    [[CONV1:%.*]] = bitcast i32* [[AA_CASTED]] to i16*
// CHECK2-NEXT:    store i16 [[TMP2]], i16* [[CONV1]], align 2
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK2-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT:    call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
// CHECK2-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
// CHECK2-NEXT:    ret void
// CHECK2:       worker.exit:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK2-NEXT:    store i16 1, i16* [[CONV]], align 4
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33
// CHECK2-SAME: (i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[AA_CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK2-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 true)
// CHECK2-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK2-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK2:       user_code.entry:
// CHECK2-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK2-NEXT:    [[TMP2:%.*]] = load i16, i16* [[CONV]], align 4
// CHECK2-NEXT:    [[CONV1:%.*]] = bitcast i32* [[AA_CASTED]] to i16*
// CHECK2-NEXT:    store i16 [[TMP2]], i16* [[CONV1]], align 2
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK2-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT:    call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
// CHECK2-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
// CHECK2-NEXT:    ret void
// CHECK2:       worker.exit:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__2
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK2-NEXT:    [[TMP0:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP1:%.*]] = bitcast i16* [[CONV]] to i8*
// CHECK2-NEXT:    store i8* [[TMP1]], i8** [[TMP0]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP3]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__3 to i8*), i8* null, i8** [[TMP4]], i32 1)
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__3
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i16* nonnull align 2 dereferenceable(2) [[AA:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[AA_ADDR:%.*]] = alloca i16*, align 4
// CHECK2-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i16* [[AA]], i16** [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP2:%.*]] = bitcast i16* [[TMP0]] to i8*
// CHECK2-NEXT:    store i8* [[TMP2]], i8** [[TMP1]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__4 to i8*), i8* null, i8** [[TMP5]], i32 1)
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__4
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i16* nonnull align 2 dereferenceable(2) [[AA:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[AA_ADDR:%.*]] = alloca i16*, align 4
// CHECK2-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i16* [[AA]], i16** [[AA_ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4
// CHECK2-NEXT:    store i16 1, i16* [[TMP0]], align 2
// CHECK2-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l23
// CHECK3-SAME: (i32 [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[A_CASTED:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK3-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK3-NEXT:    [[CONV:%.*]] = bitcast i32* [[A_ADDR]] to i8*
// CHECK3-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i1 false, i1 true, i1 true)
// CHECK3-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK3-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK3:       user_code.entry:
// CHECK3-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT:    [[TMP2:%.*]] = load i8, i8* [[CONV]], align 4
// CHECK3-NEXT:    [[CONV1:%.*]] = bitcast i32* [[A_CASTED]] to i8*
// CHECK3-NEXT:    store i8 [[TMP2]], i8* [[CONV1]], align 1
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32, i32* [[A_CASTED]], align 4
// CHECK3-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT:    call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1:[0-9]+]]
// CHECK3-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
// CHECK3-NEXT:    ret void
// CHECK3:       worker.exit:
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32 [[A:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
// CHECK3-NEXT:    [[CONV:%.*]] = bitcast i32* [[A_ADDR]] to i8*
// CHECK3-NEXT:    store i8 49, i8* [[CONV]], align 4
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l28
// CHECK3-SAME: (i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[AA_CASTED:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK3-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK3-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 false, i1 true, i1 true)
// CHECK3-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK3-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK3:       user_code.entry:
// CHECK3-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT:    [[TMP2:%.*]] = load i16, i16* [[CONV]], align 4
// CHECK3-NEXT:    [[CONV1:%.*]] = bitcast i32* [[AA_CASTED]] to i16*
// CHECK3-NEXT:    store i16 [[TMP2]], i16* [[CONV1]], align 2
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK3-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT:    call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
// CHECK3-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
// CHECK3-NEXT:    ret void
// CHECK3:       worker.exit:
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK3-NEXT:    store i16 1, i16* [[CONV]], align 4
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIcET_i_l33
// CHECK3-SAME: (i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[AA_CASTED:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK3-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK3-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 true, i1 false, i1 true)
// CHECK3-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
// CHECK3-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK3:       user_code.entry:
// CHECK3-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK3-NEXT:    [[TMP2:%.*]] = load i16, i16* [[CONV]], align 4
// CHECK3-NEXT:    [[CONV1:%.*]] = bitcast i32* [[AA_CASTED]] to i16*
// CHECK3-NEXT:    store i16 [[TMP2]], i16* [[CONV1]], align 2
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK3-NEXT:    store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT:    call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
// CHECK3-NEXT:    call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
// CHECK3-NEXT:    ret void
// CHECK3:       worker.exit:
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__2
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32 [[AA:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32 [[AA]], i32* [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16*
// CHECK3-NEXT:    [[TMP0:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK3-NEXT:    [[TMP1:%.*]] = bitcast i16* [[CONV]] to i8*
// CHECK3-NEXT:    store i8* [[TMP1]], i8** [[TMP0]], align 4
// CHECK3-NEXT:    [[TMP2:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP2]], align 4
// CHECK3-NEXT:    [[TMP4:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP3]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__3 to i8*), i8* null, i8** [[TMP4]], i32 1)
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__3
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i16* nonnull align 2 dereferenceable(2) [[AA:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[AA_ADDR:%.*]] = alloca i16*, align 4
// CHECK3-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT:    store i16* [[AA]], i16** [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK3-NEXT:    [[TMP2:%.*]] = bitcast i16* [[TMP0]] to i8*
// CHECK3-NEXT:    store i8* [[TMP2]], i8** [[TMP1]], align 4
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4
// CHECK3-NEXT:    [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__4 to i8*), i8* null, i8** [[TMP5]], i32 1)
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__4
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i16* nonnull align 2 dereferenceable(2) [[AA:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[AA_ADDR:%.*]] = alloca i16*, align 4
// CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT:    store i16* [[AA]], i16** [[AA_ADDR]], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4
// CHECK3-NEXT:    store i16 1, i16* [[TMP0]], align 2
// CHECK3-NEXT:    ret void
//