shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity
llvm-project/clang/test/OpenMP/nvptx_parallel_codegen.cpp
Shilei Tian 10068cd654 [OpenMP] Introduce kernel environment 
This patch introduces per kernel environment. Previously, flags such as execution mode are set through global variables with name like `__kernel_name_exec_mode`. They are accessible on the host by reading the corresponding global variable, but not from the device. Besides, some assumptions, such as no nested parallelism, are not per kernel basis, preventing us applying per kernel optimization in the device runtime.

This is a combination and refinement of patch series D116908, D116909, and D116910.

Depend on D155886.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D142569
2023-07-26 13:35:14 -04:00

601 lines
36 KiB
C++
Raw Blame History

// 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 -aux-triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -disable-llvm-optzns | 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 -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -aux-triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - -disable-llvm-optzns -disable-O0-optnone | FileCheck %s --check-prefix=CHECK2
// 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 if(0)
{
#pragma omp parallel
{
int a = 41;
}
a += 1;
}
#pragma omp target
{
#pragma omp parallel
{
int a = 42;
}
#pragma omp parallel if(0)
{
int a = 43;
}
#pragma omp parallel if(1)
{
int a = 44;
}
a += 1;
}
#pragma omp target if(n>40)
{
#pragma omp parallel if(n>1000)
{
int a = 45;
#pragma omp barrier
}
a += 1;
aa += 1;
b[2] += 1;
}
#pragma omp target
{
#pragma omp parallel
{
#pragma omp critical
++a;
}
++a;
}
return a;
}
int bar(int n){
int a = 0;
a += ftemplate<int>(n);
return a;
}
#endif
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26
// CHECK1-SAME: (i64 noundef [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x ptr], align 8
// CHECK1-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x ptr], align 8
// CHECK1-NEXT: [[CAPTURED_VARS_ADDRS2:%.*]] = alloca [0 x ptr], align 8
// CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_kernel_environment)
// 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(ptr @[[GLOB1:[0-9]+]])
// CHECK1-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS]], i64 0)
// CHECK1-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP1]], i32 0, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1_wrapper, ptr [[CAPTURED_VARS_ADDRS1]], i64 0)
// CHECK1-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2_wrapper, ptr [[CAPTURED_VARS_ADDRS2]], i64 0)
// CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[A_ADDR]], align 4
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP2]], 1
// CHECK1-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4
// CHECK1-NEXT: call void @__kmpc_target_deinit()
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store i32 42, ptr [[A]], align 4
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined_wrapper
// CHECK1-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[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 ptr, align 8
// CHECK1-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK1-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR3:[0-9]+]]
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store i32 43, ptr [[A]], align 4
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1_wrapper
// CHECK1-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[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 ptr, align 8
// CHECK1-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK1-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR3]]
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store i32 44, ptr [[A]], align 4
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2_wrapper
// CHECK1-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[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 ptr, align 8
// CHECK1-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK1-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR3]]
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43
// CHECK1-SAME: (i64 noundef [[N:%.*]], i64 noundef [[A:%.*]], i64 noundef [[AA:%.*]], ptr noundef nonnull align 4 dereferenceable(40) [[B:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[AA_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[B_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x ptr], align 8
// CHECK1-NEXT: store i64 [[N]], ptr [[N_ADDR]], align 8
// CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: store i64 [[AA]], ptr [[AA_ADDR]], align 8
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[B_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_kernel_environment)
// 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(ptr @[[GLOB1]])
// CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[N_ADDR]], align 4
// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1000
// CHECK1-NEXT: [[TMP4:%.*]] = zext i1 [[CMP]] to i32
// CHECK1-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP2]], i32 [[TMP4]], i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS]], i64 0)
// CHECK1-NEXT: [[TMP5:%.*]] = load i32, ptr [[A_ADDR]], align 4
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK1-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4
// CHECK1-NEXT: [[TMP6:%.*]] = load i16, ptr [[AA_ADDR]], align 2
// CHECK1-NEXT: [[CONV:%.*]] = sext i16 [[TMP6]] to i32
// CHECK1-NEXT: [[ADD1:%.*]] = add nsw i32 [[CONV]], 1
// CHECK1-NEXT: [[CONV2:%.*]] = trunc i32 [[ADD1]] to i16
// CHECK1-NEXT: store i16 [[CONV2]], ptr [[AA_ADDR]], align 2
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [10 x i32], ptr [[TMP0]], i64 0, i64 2
// CHECK1-NEXT: [[TMP7:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
// CHECK1-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK1-NEXT: store i32 [[ADD3]], ptr [[ARRAYIDX]], align 4
// CHECK1-NEXT: call void @__kmpc_target_deinit()
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store i32 45, ptr [[A]], align 4
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB2:[0-9]+]], i32 [[TMP1]])
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined_wrapper
// CHECK1-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[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 ptr, align 8
// CHECK1-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK1-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR3]]
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55
// CHECK1-SAME: (i64 noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: store i64 [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_kernel_environment)
// 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:%.*]] = load i32, ptr [[A_ADDR]], align 4
// CHECK1-NEXT: [[A1:%.*]] = call align 16 ptr @__kmpc_alloc_shared(i64 4)
// CHECK1-NEXT: store i32 [[TMP1]], ptr [[A1]], align 4
// CHECK1-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
// CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
// CHECK1-NEXT: store ptr [[A1]], ptr [[TMP3]], align 8
// CHECK1-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS]], i64 1)
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[A1]], align 4
// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[A1]], align 4
// CHECK1-NEXT: call void @__kmpc_free_shared(ptr [[A1]], i64 4)
// CHECK1-NEXT: call void @__kmpc_target_deinit()
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[CRITICAL_COUNTER:%.*]] = alloca i32, align 4
// CHECK1-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = call i64 @__kmpc_warp_active_thread_mask()
// CHECK1-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_get_hardware_thread_id_in_block()
// CHECK1-NEXT: [[NVPTX_NUM_THREADS:%.*]] = call i32 @__kmpc_get_hardware_num_threads_in_block()
// CHECK1-NEXT: store i32 0, ptr [[CRITICAL_COUNTER]], align 4
// CHECK1-NEXT: br label [[OMP_CRITICAL_LOOP:%.*]]
// CHECK1: omp.critical.loop:
// CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[CRITICAL_COUNTER]], align 4
// CHECK1-NEXT: [[TMP4:%.*]] = icmp slt i32 [[TMP3]], [[NVPTX_NUM_THREADS]]
// CHECK1-NEXT: br i1 [[TMP4]], label [[OMP_CRITICAL_TEST:%.*]], label [[OMP_CRITICAL_EXIT:%.*]]
// CHECK1: omp.critical.test:
// CHECK1-NEXT: [[TMP5:%.*]] = load i32, ptr [[CRITICAL_COUNTER]], align 4
// CHECK1-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP2]], [[TMP5]]
// CHECK1-NEXT: br i1 [[TMP6]], label [[OMP_CRITICAL_BODY:%.*]], label [[OMP_CRITICAL_SYNC:%.*]]
// CHECK1: omp.critical.body:
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP8]], ptr @"_gomp_critical_user_$var")
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP9]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[TMP0]], align 4
// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP8]], ptr @"_gomp_critical_user_$var")
// CHECK1-NEXT: br label [[OMP_CRITICAL_SYNC]]
// CHECK1: omp.critical.sync:
// CHECK1-NEXT: call void @__kmpc_syncwarp(i64 [[TMP1]])
// CHECK1-NEXT: [[TMP10:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK1-NEXT: store i32 [[TMP10]], ptr [[CRITICAL_COUNTER]], align 4
// CHECK1-NEXT: br label [[OMP_CRITICAL_LOOP]]
// CHECK1: omp.critical.exit:
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined_wrapper
// CHECK1-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[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 ptr, align 8
// CHECK1-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK1-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[GLOBAL_ARGS]], align 8
// CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 0
// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP3]], align 8
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]], ptr [[TMP4]]) #[[ATTR3]]
// CHECK1-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26
// CHECK2-SAME: (i32 noundef [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x ptr], align 4
// CHECK2-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x ptr], align 4
// CHECK2-NEXT: [[CAPTURED_VARS_ADDRS2:%.*]] = alloca [0 x ptr], align 4
// CHECK2-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_kernel_environment)
// 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(ptr @[[GLOB1:[0-9]+]])
// CHECK2-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS]], i32 0)
// CHECK2-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP1]], i32 0, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1_wrapper, ptr [[CAPTURED_VARS_ADDRS1]], i32 0)
// CHECK2-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2_wrapper, ptr [[CAPTURED_VARS_ADDRS2]], i32 0)
// CHECK2-NEXT: [[TMP2:%.*]] = load i32, ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP2]], 1
// CHECK2-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT: call void @__kmpc_target_deinit()
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK2-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT: store i32 42, ptr [[A]], align 4
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined_wrapper
// CHECK2-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR1]] {
// 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 ptr, align 4
// CHECK2-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK2-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK2-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR2:[0-9]+]]
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK2-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT: store i32 43, ptr [[A]], align 4
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1_wrapper
// CHECK2-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR1]] {
// 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 ptr, align 4
// CHECK2-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK2-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK2-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined1(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR2]]
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK2-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT: store i32 44, ptr [[A]], align 4
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2_wrapper
// CHECK2-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR1]] {
// 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 ptr, align 4
// CHECK2-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK2-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK2-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l26_omp_outlined2(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR2]]
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43
// CHECK2-SAME: (i32 noundef [[N:%.*]], i32 noundef [[A:%.*]], i32 noundef [[AA:%.*]], ptr noundef nonnull align 4 dereferenceable(40) [[B:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT: [[AA_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT: [[B_ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x ptr], align 4
// CHECK2-NEXT: store i32 [[N]], ptr [[N_ADDR]], align 4
// CHECK2-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT: store i32 [[AA]], ptr [[AA_ADDR]], align 4
// CHECK2-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 4
// CHECK2-NEXT: [[TMP0:%.*]] = load ptr, ptr [[B_ADDR]], align 4
// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_kernel_environment)
// 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(ptr @[[GLOB1]])
// CHECK2-NEXT: [[TMP3:%.*]] = load i32, ptr [[N_ADDR]], align 4
// CHECK2-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1000
// CHECK2-NEXT: [[TMP4:%.*]] = zext i1 [[CMP]] to i32
// CHECK2-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP2]], i32 [[TMP4]], i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS]], i32 0)
// CHECK2-NEXT: [[TMP5:%.*]] = load i32, ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK2-NEXT: store i32 [[ADD]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[TMP6:%.*]] = load i16, ptr [[AA_ADDR]], align 2
// CHECK2-NEXT: [[CONV:%.*]] = sext i16 [[TMP6]] to i32
// CHECK2-NEXT: [[ADD1:%.*]] = add nsw i32 [[CONV]], 1
// CHECK2-NEXT: [[CONV2:%.*]] = trunc i32 [[ADD1]] to i16
// CHECK2-NEXT: store i16 [[CONV2]], ptr [[AA_ADDR]], align 2
// CHECK2-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [10 x i32], ptr [[TMP0]], i32 0, i32 2
// CHECK2-NEXT: [[TMP7:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
// CHECK2-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK2-NEXT: store i32 [[ADD3]], ptr [[ARRAYIDX]], align 4
// CHECK2-NEXT: call void @__kmpc_target_deinit()
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK2-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT: store i32 45, ptr [[A]], align 4
// CHECK2-NEXT: [[TMP0:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: [[TMP1:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK2-NEXT: call void @__kmpc_barrier(ptr @[[GLOB2:[0-9]+]], i32 [[TMP1]])
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined_wrapper
// CHECK2-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR1]] {
// 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 ptr, align 4
// CHECK2-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK2-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK2-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l43_omp_outlined(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]]) #[[ATTR2]]
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55
// CHECK2-SAME: (i32 noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x ptr], align 4
// CHECK2-NEXT: store i32 [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_kernel_environment)
// 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:%.*]] = load i32, ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[A1:%.*]] = call align 4 ptr @__kmpc_alloc_shared(i32 4)
// CHECK2-NEXT: store i32 [[TMP1]], ptr [[A1]], align 4
// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1]])
// CHECK2-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK2-NEXT: store ptr [[A1]], ptr [[TMP3]], align 4
// CHECK2-NEXT: call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS]], i32 1)
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, ptr [[A1]], align 4
// CHECK2-NEXT: [[INC:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK2-NEXT: store i32 [[INC]], ptr [[A1]], align 4
// CHECK2-NEXT: call void @__kmpc_free_shared(ptr [[A1]], i32 4)
// CHECK2-NEXT: call void @__kmpc_target_deinit()
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT: [[CRITICAL_COUNTER:%.*]] = alloca i32, align 4
// CHECK2-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 4
// CHECK2-NEXT: [[TMP1:%.*]] = call i64 @__kmpc_warp_active_thread_mask()
// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_get_hardware_thread_id_in_block()
// CHECK2-NEXT: [[NVPTX_NUM_THREADS:%.*]] = call i32 @__kmpc_get_hardware_num_threads_in_block()
// CHECK2-NEXT: store i32 0, ptr [[CRITICAL_COUNTER]], align 4
// CHECK2-NEXT: br label [[OMP_CRITICAL_LOOP:%.*]]
// CHECK2: omp.critical.loop:
// CHECK2-NEXT: [[TMP3:%.*]] = load i32, ptr [[CRITICAL_COUNTER]], align 4
// CHECK2-NEXT: [[TMP4:%.*]] = icmp slt i32 [[TMP3]], [[NVPTX_NUM_THREADS]]
// CHECK2-NEXT: br i1 [[TMP4]], label [[OMP_CRITICAL_TEST:%.*]], label [[OMP_CRITICAL_EXIT:%.*]]
// CHECK2: omp.critical.test:
// CHECK2-NEXT: [[TMP5:%.*]] = load i32, ptr [[CRITICAL_COUNTER]], align 4
// CHECK2-NEXT: [[TMP6:%.*]] = icmp eq i32 [[TMP2]], [[TMP5]]
// CHECK2-NEXT: br i1 [[TMP6]], label [[OMP_CRITICAL_BODY:%.*]], label [[OMP_CRITICAL_SYNC:%.*]]
// CHECK2: omp.critical.body:
// CHECK2-NEXT: [[TMP7:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK2-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP8]], ptr @"_gomp_critical_user_$var")
// CHECK2-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK2-NEXT: [[INC:%.*]] = add nsw i32 [[TMP9]], 1
// CHECK2-NEXT: store i32 [[INC]], ptr [[TMP0]], align 4
// CHECK2-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP8]], ptr @"_gomp_critical_user_$var")
// CHECK2-NEXT: br label [[OMP_CRITICAL_SYNC]]
// CHECK2: omp.critical.sync:
// CHECK2-NEXT: call void @__kmpc_syncwarp(i64 [[TMP1]])
// CHECK2-NEXT: [[TMP10:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK2-NEXT: store i32 [[TMP10]], ptr [[CRITICAL_COUNTER]], align 4
// CHECK2-NEXT: br label [[OMP_CRITICAL_LOOP]]
// CHECK2: omp.critical.exit:
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined_wrapper
// CHECK2-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR1]] {
// 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 ptr, align 4
// CHECK2-NEXT: store i16 [[TMP0]], ptr [[DOTADDR]], align 2
// CHECK2-NEXT: store i32 [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK2-NEXT: store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT: call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS]])
// CHECK2-NEXT: [[TMP2:%.*]] = load ptr, ptr [[GLOBAL_ARGS]], align 4
// CHECK2-NEXT: [[TMP3:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i32 0
// CHECK2-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP3]], align 4
// CHECK2-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l55_omp_outlined(ptr [[DOTADDR1]], ptr [[DOTZERO_ADDR]], ptr [[TMP4]]) #[[ATTR2]]
// CHECK2-NEXT: ret void
//
Reference in New Issue View Git Blame Copy Permalink