xref: /llvm-project-15.0.7/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp (revision ac90dfc4)

// 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 -fopenmp-cuda-mode -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 -fopenmp-cuda-mode -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 - -disable-llvm-optzns | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-cuda-mode -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 -fopenmp-cuda-mode -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 - -disable-llvm-optzns | FileCheck %s --check-prefix=CHECK2
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-cuda-mode -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 - -disable-llvm-optzns | FileCheck %s --check-prefix=CHECK3

// expected-no-diagnostics
#ifndef HEADER
#define HEADER

void work(int *C) {
  #pragma omp atomic
  ++(*C);
}

void use(int *C) {
  #pragma omp parallel num_threads(2)
  work(C);
}

int main() {
  int C = 0;
  #pragma omp target map(C)
  {
    use(&C);
    #pragma omp parallel num_threads(2)
    use(&C);
  }

  return C;
}

#endif
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l25
// CHECK1-SAME: (i32* nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8
// CHECK1-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = 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 [[TMP1]], -1
// CHECK1-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK1:       user_code.entry:
// CHECK1-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT:    call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5:[0-9]+]]
// CHECK1-NEXT:    call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 2)
// CHECK1-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK1-NEXT:    store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP5]], i64 1)
// 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 {{[^@]+}}@_Z3usePi
// CHECK1-SAME: (i32* [[C:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK1-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 2)
// CHECK1-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP2:%.*]] = bitcast i32** [[C_ADDR]] to i8*
// CHECK1-NEXT:    store i8* [[TMP2]], i8** [[TMP1]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32**)* @__omp_outlined__1 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__1_wrapper to i8*), i8** [[TMP3]], i64 1)
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32*, 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* [[C]], i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined___wrapper
// CHECK1-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT:    store i16 [[TMP0]], i16* [[DOTADDR]], align 2
// CHECK1-NEXT:    store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
// CHECK1-NEXT:    call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
// CHECK1-NEXT:    [[TMP2:%.*]] = load i8**, i8*** [[GLOBAL_ARGS]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i64 0
// CHECK1-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32**
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 8
// CHECK1-NEXT:    call void @__omp_outlined__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR3:[0-9]+]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK1-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32** nonnull align 8 dereferenceable(8) [[C:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32**, 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** [[C]], i32*** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32**, i32*** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[TMP0]], align 8
// CHECK1-NEXT:    call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR5]]
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_Z4workPi
// CHECK1-SAME: (i32* [[C:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[ATOMIC_TEMP1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK1-NEXT:    [[TMP2:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8*
// CHECK1-NEXT:    call void @__atomic_load(i64 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR5]]
// CHECK1-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK1:       atomic_cont:
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, i32* [[ATOMIC_TEMP]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP3]], 1
// CHECK1-NEXT:    store i32 [[ADD]], i32* [[ATOMIC_TEMP1]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK1-NEXT:    [[TMP5:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8*
// CHECK1-NEXT:    [[TMP6:%.*]] = bitcast i32* [[ATOMIC_TEMP1]] to i8*
// CHECK1-NEXT:    [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i64 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR5]]
// CHECK1-NEXT:    br i1 [[CALL]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK1:       atomic_exit:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__1_wrapper
// CHECK1-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK1-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca i8**, align 8
// CHECK1-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT:    store i16 [[TMP0]], i16* [[DOTADDR]], align 2
// CHECK1-NEXT:    store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
// CHECK1-NEXT:    call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
// CHECK1-NEXT:    [[TMP2:%.*]] = load i8**, i8*** [[GLOBAL_ARGS]], align 8
// CHECK1-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i64 0
// CHECK1-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32***
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32**, i32*** [[TMP4]], align 8
// CHECK1-NEXT:    call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32** [[TMP5]]) #[[ATTR3]]
// CHECK1-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l25
// CHECK2-SAME: (i32* nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK2-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = 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 [[TMP1]], -1
// CHECK2-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK2:       user_code.entry:
// CHECK2-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT:    call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5:[0-9]+]]
// CHECK2-NEXT:    call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 2)
// CHECK2-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK2-NEXT:    store i8* [[TMP4]], i8** [[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* @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP5]], i32 1)
// 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 {{[^@]+}}@_Z3usePi
// CHECK2-SAME: (i32* [[C:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK2-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 2)
// CHECK2-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP2:%.*]] = bitcast i32** [[C_ADDR]] to i8*
// CHECK2-NEXT:    store i8* [[TMP2]], i8** [[TMP1]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32**)* @__omp_outlined__1 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__1_wrapper to i8*), i8** [[TMP3]], i32 1)
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[C_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* [[C]], i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined___wrapper
// CHECK2-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca i8**, align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT:    store i16 [[TMP0]], i16* [[DOTADDR]], align 2
// CHECK2-NEXT:    store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
// CHECK2-NEXT:    call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
// CHECK2-NEXT:    [[TMP2:%.*]] = load i8**, i8*** [[GLOBAL_ARGS]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i32 0
// CHECK2-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32**
// CHECK2-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 4
// CHECK2-NEXT:    call void @__omp_outlined__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR3:[0-9]+]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK2-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32** nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[C_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** [[C]], i32*** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32**, i32*** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[TMP0]], align 4
// CHECK2-NEXT:    call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR5]]
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@_Z4workPi
// CHECK2-SAME: (i32* [[C:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 4
// CHECK2-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[ATOMIC_TEMP1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK2-NEXT:    [[TMP2:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8*
// CHECK2-NEXT:    call void @__atomic_load(i32 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR5]]
// CHECK2-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK2:       atomic_cont:
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[ATOMIC_TEMP]], align 4
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP3]], 1
// CHECK2-NEXT:    store i32 [[ADD]], i32* [[ATOMIC_TEMP1]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK2-NEXT:    [[TMP5:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8*
// CHECK2-NEXT:    [[TMP6:%.*]] = bitcast i32* [[ATOMIC_TEMP1]] to i8*
// CHECK2-NEXT:    [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i32 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR5]]
// CHECK2-NEXT:    br i1 [[CALL]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK2:       atomic_exit:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__1_wrapper
// CHECK2-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK2-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca i8**, align 4
// CHECK2-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT:    store i16 [[TMP0]], i16* [[DOTADDR]], align 2
// CHECK2-NEXT:    store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
// CHECK2-NEXT:    call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
// CHECK2-NEXT:    [[TMP2:%.*]] = load i8**, i8*** [[GLOBAL_ARGS]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i32 0
// CHECK2-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32***
// CHECK2-NEXT:    [[TMP5:%.*]] = load i32**, i32*** [[TMP4]], align 4
// CHECK2-NEXT:    call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32** [[TMP5]]) #[[ATTR3]]
// CHECK2-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l25
// CHECK3-SAME: (i32* nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK3-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP1:%.*]] = 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 [[TMP1]], -1
// CHECK3-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
// CHECK3:       user_code.entry:
// CHECK3-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT:    call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5:[0-9]+]]
// CHECK3-NEXT:    call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 2)
// CHECK3-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK3-NEXT:    [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK3-NEXT:    store i8* [[TMP4]], i8** [[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* @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP5]], i32 1)
// 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 {{[^@]+}}@_Z3usePi
// CHECK3-SAME: (i32* [[C:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK3-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 2)
// CHECK3-NEXT:    [[TMP1:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK3-NEXT:    [[TMP2:%.*]] = bitcast i32** [[C_ADDR]] to i8*
// CHECK3-NEXT:    store i8* [[TMP2]], i8** [[TMP1]], align 4
// CHECK3-NEXT:    [[TMP3:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT:    call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32**)* @__omp_outlined__1 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__1_wrapper to i8*), i8** [[TMP3]], i32 1)
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[C_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* [[C]], i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined___wrapper
// CHECK3-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2:[0-9]+]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK3-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca i8**, align 4
// CHECK3-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK3-NEXT:    store i16 [[TMP0]], i16* [[DOTADDR]], align 2
// CHECK3-NEXT:    store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
// CHECK3-NEXT:    call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
// CHECK3-NEXT:    [[TMP2:%.*]] = load i8**, i8*** [[GLOBAL_ARGS]], align 4
// CHECK3-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i32 0
// CHECK3-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32**
// CHECK3-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 4
// CHECK3-NEXT:    call void @__omp_outlined__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR3:[0-9]+]]
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK3-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]], i32** nonnull align 4 dereferenceable(4) [[C:%.*]]) #[[ATTR0]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[C_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** [[C]], i32*** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32**, i32*** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[TMP0]], align 4
// CHECK3-NEXT:    call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR5]]
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_Z4workPi
// CHECK3-SAME: (i32* [[C:%.*]]) #[[ATTR1]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[C_ADDR:%.*]] = alloca i32*, align 4
// CHECK3-NEXT:    [[ATOMIC_TEMP:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[ATOMIC_TEMP1:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    store i32* [[C]], i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4
// CHECK3-NEXT:    [[TMP1:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK3-NEXT:    [[TMP2:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8*
// CHECK3-NEXT:    call void @__atomic_load(i32 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR5]]
// CHECK3-NEXT:    br label [[ATOMIC_CONT:%.*]]
// CHECK3:       atomic_cont:
// CHECK3-NEXT:    [[TMP3:%.*]] = load i32, i32* [[ATOMIC_TEMP]], align 4
// CHECK3-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP3]], 1
// CHECK3-NEXT:    store i32 [[ADD]], i32* [[ATOMIC_TEMP1]], align 4
// CHECK3-NEXT:    [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8*
// CHECK3-NEXT:    [[TMP5:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8*
// CHECK3-NEXT:    [[TMP6:%.*]] = bitcast i32* [[ATOMIC_TEMP1]] to i8*
// CHECK3-NEXT:    [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i32 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR5]]
// CHECK3-NEXT:    br i1 [[CALL]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]]
// CHECK3:       atomic_exit:
// CHECK3-NEXT:    ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__1_wrapper
// CHECK3-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2]] {
// CHECK3-NEXT:  entry:
// CHECK3-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK3-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca i8**, align 4
// CHECK3-NEXT:    store i32 0, i32* [[DOTZERO_ADDR]], align 4
// CHECK3-NEXT:    store i16 [[TMP0]], i16* [[DOTADDR]], align 2
// CHECK3-NEXT:    store i32 [[TMP1]], i32* [[DOTADDR1]], align 4
// CHECK3-NEXT:    call void @__kmpc_get_shared_variables(i8*** [[GLOBAL_ARGS]])
// CHECK3-NEXT:    [[TMP2:%.*]] = load i8**, i8*** [[GLOBAL_ARGS]], align 4
// CHECK3-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i32 0
// CHECK3-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32***
// CHECK3-NEXT:    [[TMP5:%.*]] = load i32**, i32*** [[TMP4]], align 4
// CHECK3-NEXT:    call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32** [[TMP5]]) #[[ATTR3]]
// CHECK3-NEXT:    ret void
//