caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
dda73336ad22bd0b5ecda17040c50fb10fcbe5fb
clang-p2996/clang/test/OpenMP/teams_generic_loop_reduction_codegen.cpp
David Pagan a12836647e [OpenMP][CodeGen] Improved codegen for combined loop directives (#87278) 
IR for 'target teams loop' is now dependent on suitability of associated
loop-nest.

If a loop-nest:

- does not contain a function call, or
- the -fopenmp-assume-no-nested-parallelism has been specified,
- or the call is to an OpenMP API AND
- does not contain nested loop bind(parallel) directives

then it can be emitted as 'target teams distribute parallel for', which
is the current default. Otherwise, it is emitted as 'target teams
distribute'.

Added debug output indicating how 'target teams loop' was emitted. Flag
is -mllvm -debug-only=target-teams-loop-codegen

Added LIT tests explicitly verifying 'target teams loop' emitted as a
parallel loop and a distribute loop.

Updated other 'loop' related tests as needed to reflect change in IR.
- These updates account for most of the changed files and
additions/deletions.
2024-04-10 13:09:17 -07:00

937 lines
62 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 _
// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK3
// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK3
// RUN: %clang_cc1 -DCHECK -verify -fopenmp-simd -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -DCHECK -fopenmp-simd -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -DCHECK -fopenmp-simd -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -DCHECK -verify -fopenmp-simd -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -DCHECK -fopenmp-simd -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -DCHECK -fopenmp-simd -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK9
// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix=CHECK9
// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp-simd -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// RUN: %clang_cc1 -DLAMBDA -fopenmp-simd -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -DLAMBDA -fopenmp-simd -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
// expected-no-diagnostics
#ifndef HEADER
#define HEADER
template <typename T>
T tmain() {
T t_var = T();
T vec[] = {1, 2};
#pragma omp target
#pragma omp teams loop reduction(+: t_var)
for (int i = 0; i < 2; ++i) {
t_var += (T) i;
}
return T();
}
int main() {
static int sivar;
#ifdef LAMBDA
[&]() {
#pragma omp target
#pragma omp teams loop reduction(+: sivar)
for (int i = 0; i < 2; ++i) {
// Skip global and bound tid vars
// Skip global and bound tid vars, and prev lb and ub vars
// skip loop vars
sivar += i;
[&]() {
sivar += 4;
}();
}
}();
return 0;
#else
#pragma omp target
#pragma omp teams loop reduction(+: sivar)
for (int i = 0; i < 2; ++i) {
sivar += i;
}
return tmain<int>();
#endif
}
// Skip global and bound tid vars
// Skip global and bound tid vars, and prev lb and ub
// skip loop vars
// Skip global and bound tid vars
// Skip global and bound tid vars, and prev lb and ub vars
// skip loop vars
#endif
// CHECK1-LABEL: define {{[^@]+}}@main
// CHECK1-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[SIVAR_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8
// CHECK1-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr @_ZZ4mainE5sivar, align 4
// CHECK1-NEXT: store i32 [[TMP0]], ptr [[SIVAR_CASTED]], align 4
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[SIVAR_CASTED]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK1-NEXT: store i64 [[TMP1]], ptr [[TMP2]], align 8
// CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK1-NEXT: store i64 [[TMP1]], ptr [[TMP3]], align 8
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0
// CHECK1-NEXT: store ptr null, ptr [[TMP4]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0
// CHECK1-NEXT: store i32 3, ptr [[TMP7]], align 4
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1
// CHECK1-NEXT: store i32 1, ptr [[TMP8]], align 4
// CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2
// CHECK1-NEXT: store ptr [[TMP5]], ptr [[TMP9]], align 8
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP10]], align 8
// CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4
// CHECK1-NEXT: store ptr @.offload_sizes, ptr [[TMP11]], align 8
// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5
// CHECK1-NEXT: store ptr @.offload_maptypes, ptr [[TMP12]], align 8
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6
// CHECK1-NEXT: store ptr null, ptr [[TMP13]], align 8
// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7
// CHECK1-NEXT: store ptr null, ptr [[TMP14]], align 8
// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8
// CHECK1-NEXT: store i64 2, ptr [[TMP15]], align 8
// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9
// CHECK1-NEXT: store i64 0, ptr [[TMP16]], align 8
// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10
// CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP17]], align 4
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11
// CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP18]], align 4
// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12
// CHECK1-NEXT: store i32 0, ptr [[TMP19]], align 4
// CHECK1-NEXT: [[TMP20:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB3:[0-9]+]], i64 -1, i32 0, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.region_id, ptr [[KERNEL_ARGS]])
// CHECK1-NEXT: [[TMP21:%.*]] = icmp ne i32 [[TMP20]], 0
// CHECK1-NEXT: br i1 [[TMP21]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]]
// CHECK1: omp_offload.failed:
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68(i64 [[TMP1]]) #[[ATTR2:[0-9]+]]
// CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]]
// CHECK1: omp_offload.cont:
// CHECK1-NEXT: [[CALL:%.*]] = call noundef signext i32 @_Z5tmainIiET_v()
// CHECK1-NEXT: ret i32 [[CALL]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68
// CHECK1-SAME: (i64 noundef [[SIVAR:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[SIVAR_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: store i64 [[SIVAR]], ptr [[SIVAR_ADDR]], align 8
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB3]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined, ptr [[SIVAR_ADDR]])
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[SIVAR:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[SIVAR_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[SIVAR1:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[I:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x ptr], align 8
// 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 [[SIVAR]], ptr [[SIVAR_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[SIVAR_ADDR]], align 8
// CHECK1-NEXT: store i32 0, ptr [[SIVAR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1:[0-9]+]], i32 [[TMP2]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
// CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1
// CHECK1-NEXT: br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1: cond.true:
// CHECK1-NEXT: br label [[COND_END:%.*]]
// CHECK1: cond.false:
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 1, [[COND_TRUE]] ], [ [[TMP4]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT: store i32 [[TMP5]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
// CHECK1-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: [[CMP2:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK1-NEXT: br i1 [[CMP2]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP8]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], align 4
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[TMP10:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK1-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP10]], [[TMP9]]
// CHECK1-NEXT: store i32 [[ADD3]], ptr [[SIVAR1]], align 4
// CHECK1-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK1: omp.body.continue:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1: omp.inner.for.inc:
// CHECK1-NEXT: [[TMP11:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP11]], 1
// CHECK1-NEXT: store i32 [[ADD4]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK1: omp.inner.for.end:
// CHECK1-NEXT: br label [[OMP_LOOP_EXIT:%.*]]
// CHECK1: omp.loop.exit:
// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP2]])
// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: store ptr [[SIVAR1]], ptr [[TMP12]], align 8
// CHECK1-NEXT: [[TMP13:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP2]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: switch i32 [[TMP13]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK1-NEXT: i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK1-NEXT: i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK1-NEXT: ]
// CHECK1: .omp.reduction.case1:
// CHECK1-NEXT: [[TMP14:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK1-NEXT: [[TMP15:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK1-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP14]], [[TMP15]]
// CHECK1-NEXT: store i32 [[ADD5]], ptr [[TMP0]], align 4
// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK1-NEXT: [[TMP17:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP16]] monotonic, align 4
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined.omp.reduction.reduction_func
// CHECK1-SAME: (ptr noundef [[TMP0:%.*]], ptr noundef [[TMP1:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP4]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP2]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP8]], [[TMP9]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP7]], align 4
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@_Z5tmainIiET_v
// CHECK1-SAME: () #[[ATTR5:[0-9]+]] comdat {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[T_VAR:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[VEC:%.*]] = alloca [2 x i32], align 4
// CHECK1-NEXT: [[T_VAR_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x ptr], align 8
// CHECK1-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8
// CHECK1-NEXT: store i32 0, ptr [[T_VAR]], align 4
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i64 8, i1 false)
// CHECK1-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4
// CHECK1-NEXT: store i32 [[TMP0]], ptr [[T_VAR_CASTED]], align 4
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[T_VAR_CASTED]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK1-NEXT: store i64 [[TMP1]], ptr [[TMP2]], align 8
// CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK1-NEXT: store i64 [[TMP1]], ptr [[TMP3]], align 8
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0
// CHECK1-NEXT: store ptr null, ptr [[TMP4]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0
// CHECK1-NEXT: store i32 3, ptr [[TMP7]], align 4
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1
// CHECK1-NEXT: store i32 1, ptr [[TMP8]], align 4
// CHECK1-NEXT: [[TMP9:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2
// CHECK1-NEXT: store ptr [[TMP5]], ptr [[TMP9]], align 8
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[TMP10]], align 8
// CHECK1-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4
// CHECK1-NEXT: store ptr @.offload_sizes.1, ptr [[TMP11]], align 8
// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5
// CHECK1-NEXT: store ptr @.offload_maptypes.2, ptr [[TMP12]], align 8
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6
// CHECK1-NEXT: store ptr null, ptr [[TMP13]], align 8
// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7
// CHECK1-NEXT: store ptr null, ptr [[TMP14]], align 8
// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8
// CHECK1-NEXT: store i64 2, ptr [[TMP15]], align 8
// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9
// CHECK1-NEXT: store i64 0, ptr [[TMP16]], align 8
// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10
// CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP17]], align 4
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11
// CHECK1-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP18]], align 4
// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12
// CHECK1-NEXT: store i32 0, ptr [[TMP19]], align 4
// CHECK1-NEXT: [[TMP20:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB3]], i64 -1, i32 0, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.region_id, ptr [[KERNEL_ARGS]])
// CHECK1-NEXT: [[TMP21:%.*]] = icmp ne i32 [[TMP20]], 0
// CHECK1-NEXT: br i1 [[TMP21]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]]
// CHECK1: omp_offload.failed:
// CHECK1-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32(i64 [[TMP1]]) #[[ATTR2]]
// CHECK1-NEXT: br label [[OMP_OFFLOAD_CONT]]
// CHECK1: omp_offload.cont:
// CHECK1-NEXT: ret i32 0
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32
// CHECK1-SAME: (i64 noundef [[T_VAR:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[T_VAR_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: store i64 [[T_VAR]], ptr [[T_VAR_ADDR]], align 8
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB3]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined, ptr [[T_VAR_ADDR]])
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[T_VAR:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[T_VAR_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[T_VAR1:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[I:%.*]] = alloca i32, align 4
// CHECK1-NEXT: [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x ptr], align 8
// 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 [[T_VAR]], ptr [[T_VAR_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8
// CHECK1-NEXT: store i32 0, ptr [[T_VAR1]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP2]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
// CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1
// CHECK1-NEXT: br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1: cond.true:
// CHECK1-NEXT: br label [[COND_END:%.*]]
// CHECK1: cond.false:
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 1, [[COND_TRUE]] ], [ [[TMP4]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT: store i32 [[TMP5]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
// CHECK1-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT: [[CMP2:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK1-NEXT: br i1 [[CMP2]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP8]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], align 4
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[TMP10:%.*]] = load i32, ptr [[T_VAR1]], align 4
// CHECK1-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP10]], [[TMP9]]
// CHECK1-NEXT: store i32 [[ADD3]], ptr [[T_VAR1]], align 4
// CHECK1-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK1: omp.body.continue:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1: omp.inner.for.inc:
// CHECK1-NEXT: [[TMP11:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP11]], 1
// CHECK1-NEXT: store i32 [[ADD4]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK1: omp.inner.for.end:
// CHECK1-NEXT: br label [[OMP_LOOP_EXIT:%.*]]
// CHECK1: omp.loop.exit:
// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP2]])
// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[TMP12]], align 8
// CHECK1-NEXT: [[TMP13:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: switch i32 [[TMP13]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK1-NEXT: i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK1-NEXT: i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK1-NEXT: ]
// CHECK1: .omp.reduction.case1:
// CHECK1-NEXT: [[TMP14:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK1-NEXT: [[TMP15:%.*]] = load i32, ptr [[T_VAR1]], align 4
// CHECK1-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP14]], [[TMP15]]
// CHECK1-NEXT: store i32 [[ADD5]], ptr [[TMP0]], align 4
// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[T_VAR1]], align 4
// CHECK1-NEXT: [[TMP17:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP16]] monotonic, align 4
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
// CHECK1-NEXT: ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined.omp.reduction.reduction_func
// CHECK1-SAME: (ptr noundef [[TMP0:%.*]], ptr noundef [[TMP1:%.*]]) #[[ATTR3]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP4]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP2]], i64 0, i64 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK1-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP8]], [[TMP9]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP7]], align 4
// CHECK1-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@main
// CHECK3-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[SIVAR_CASTED:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8
// CHECK3-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr @_ZZ4mainE5sivar, align 4
// CHECK3-NEXT: store i32 [[TMP0]], ptr [[SIVAR_CASTED]], align 4
// CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[SIVAR_CASTED]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK3-NEXT: store i32 [[TMP1]], ptr [[TMP2]], align 4
// CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK3-NEXT: store i32 [[TMP1]], ptr [[TMP3]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0
// CHECK3-NEXT: store ptr null, ptr [[TMP4]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0
// CHECK3-NEXT: store i32 3, ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1
// CHECK3-NEXT: store i32 1, ptr [[TMP8]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2
// CHECK3-NEXT: store ptr [[TMP5]], ptr [[TMP9]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP10]], align 4
// CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4
// CHECK3-NEXT: store ptr @.offload_sizes, ptr [[TMP11]], align 4
// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5
// CHECK3-NEXT: store ptr @.offload_maptypes, ptr [[TMP12]], align 4
// CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6
// CHECK3-NEXT: store ptr null, ptr [[TMP13]], align 4
// CHECK3-NEXT: [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7
// CHECK3-NEXT: store ptr null, ptr [[TMP14]], align 4
// CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8
// CHECK3-NEXT: store i64 2, ptr [[TMP15]], align 8
// CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9
// CHECK3-NEXT: store i64 0, ptr [[TMP16]], align 8
// CHECK3-NEXT: [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10
// CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP17]], align 4
// CHECK3-NEXT: [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11
// CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP18]], align 4
// CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12
// CHECK3-NEXT: store i32 0, ptr [[TMP19]], align 4
// CHECK3-NEXT: [[TMP20:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB3:[0-9]+]], i64 -1, i32 0, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.region_id, ptr [[KERNEL_ARGS]])
// CHECK3-NEXT: [[TMP21:%.*]] = icmp ne i32 [[TMP20]], 0
// CHECK3-NEXT: br i1 [[TMP21]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]]
// CHECK3: omp_offload.failed:
// CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68(i32 [[TMP1]]) #[[ATTR2:[0-9]+]]
// CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]]
// CHECK3: omp_offload.cont:
// CHECK3-NEXT: [[CALL:%.*]] = call noundef i32 @_Z5tmainIiET_v()
// CHECK3-NEXT: ret i32 [[CALL]]
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68
// CHECK3-SAME: (i32 noundef [[SIVAR:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[SIVAR_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT: store i32 [[SIVAR]], ptr [[SIVAR_ADDR]], align 4
// CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB3]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined, ptr [[SIVAR_ADDR]])
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined
// CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[SIVAR:%.*]]) #[[ATTR1]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[SIVAR_ADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[SIVAR1:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[I:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT: store ptr [[SIVAR]], ptr [[SIVAR_ADDR]], align 4
// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[SIVAR_ADDR]], align 4
// CHECK3-NEXT: store i32 0, ptr [[SIVAR1]], align 4
// CHECK3-NEXT: store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK3-NEXT: store i32 1, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK3-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
// CHECK3-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1:[0-9]+]], i32 [[TMP2]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
// CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1
// CHECK3-NEXT: br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK3: cond.true:
// CHECK3-NEXT: br label [[COND_END:%.*]]
// CHECK3: cond.false:
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: br label [[COND_END]]
// CHECK3: cond.end:
// CHECK3-NEXT: [[COND:%.*]] = phi i32 [ 1, [[COND_TRUE]] ], [ [[TMP4]], [[COND_FALSE]] ]
// CHECK3-NEXT: store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK3-NEXT: store i32 [[TMP5]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3: omp.inner.for.cond:
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: [[CMP2:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK3-NEXT: br i1 [[CMP2]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK3: omp.inner.for.body:
// CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP8]], 1
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[I]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK3-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP10]], [[TMP9]]
// CHECK3-NEXT: store i32 [[ADD3]], ptr [[SIVAR1]], align 4
// CHECK3-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK3: omp.body.continue:
// CHECK3-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK3: omp.inner.for.inc:
// CHECK3-NEXT: [[TMP11:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP11]], 1
// CHECK3-NEXT: store i32 [[ADD4]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK3: omp.inner.for.end:
// CHECK3-NEXT: br label [[OMP_LOOP_EXIT:%.*]]
// CHECK3: omp.loop.exit:
// CHECK3-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP2]])
// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: store ptr [[SIVAR1]], ptr [[TMP12]], align 4
// CHECK3-NEXT: [[TMP13:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP2]], i32 1, i32 4, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT: switch i32 [[TMP13]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK3-NEXT: i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK3-NEXT: i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK3-NEXT: ]
// CHECK3: .omp.reduction.case1:
// CHECK3-NEXT: [[TMP14:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK3-NEXT: [[TMP15:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK3-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP14]], [[TMP15]]
// CHECK3-NEXT: store i32 [[ADD5]], ptr [[TMP0]], align 4
// CHECK3-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], ptr @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3: .omp.reduction.case2:
// CHECK3-NEXT: [[TMP16:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK3-NEXT: [[TMP17:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP16]] monotonic, align 4
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3: .omp.reduction.default:
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l68.omp_outlined.omp.reduction.reduction_func
// CHECK3-SAME: (ptr noundef [[TMP0:%.*]], ptr noundef [[TMP1:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[DOTADDR1:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 4
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 4
// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP4]], align 4
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP2]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP8]], [[TMP9]]
// CHECK3-NEXT: store i32 [[ADD]], ptr [[TMP7]], align 4
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@_Z5tmainIiET_v
// CHECK3-SAME: () #[[ATTR5:[0-9]+]] comdat {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[T_VAR:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[VEC:%.*]] = alloca [2 x i32], align 4
// CHECK3-NEXT: [[T_VAR_CASTED:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8
// CHECK3-NEXT: store i32 0, ptr [[T_VAR]], align 4
// CHECK3-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[VEC]], ptr align 4 @__const._Z5tmainIiET_v.vec, i32 8, i1 false)
// CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[T_VAR]], align 4
// CHECK3-NEXT: store i32 [[TMP0]], ptr [[T_VAR_CASTED]], align 4
// CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[T_VAR_CASTED]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK3-NEXT: store i32 [[TMP1]], ptr [[TMP2]], align 4
// CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK3-NEXT: store i32 [[TMP1]], ptr [[TMP3]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0
// CHECK3-NEXT: store ptr null, ptr [[TMP4]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOFFLOAD_PTRS]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 0
// CHECK3-NEXT: store i32 3, ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 1
// CHECK3-NEXT: store i32 1, ptr [[TMP8]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 2
// CHECK3-NEXT: store ptr [[TMP5]], ptr [[TMP9]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 3
// CHECK3-NEXT: store ptr [[TMP6]], ptr [[TMP10]], align 4
// CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 4
// CHECK3-NEXT: store ptr @.offload_sizes.1, ptr [[TMP11]], align 4
// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 5
// CHECK3-NEXT: store ptr @.offload_maptypes.2, ptr [[TMP12]], align 4
// CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 6
// CHECK3-NEXT: store ptr null, ptr [[TMP13]], align 4
// CHECK3-NEXT: [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 7
// CHECK3-NEXT: store ptr null, ptr [[TMP14]], align 4
// CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 8
// CHECK3-NEXT: store i64 2, ptr [[TMP15]], align 8
// CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 9
// CHECK3-NEXT: store i64 0, ptr [[TMP16]], align 8
// CHECK3-NEXT: [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 10
// CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP17]], align 4
// CHECK3-NEXT: [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 11
// CHECK3-NEXT: store [3 x i32] zeroinitializer, ptr [[TMP18]], align 4
// CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], ptr [[KERNEL_ARGS]], i32 0, i32 12
// CHECK3-NEXT: store i32 0, ptr [[TMP19]], align 4
// CHECK3-NEXT: [[TMP20:%.*]] = call i32 @__tgt_target_kernel(ptr @[[GLOB3]], i64 -1, i32 0, i32 0, ptr @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.region_id, ptr [[KERNEL_ARGS]])
// CHECK3-NEXT: [[TMP21:%.*]] = icmp ne i32 [[TMP20]], 0
// CHECK3-NEXT: br i1 [[TMP21]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]]
// CHECK3: omp_offload.failed:
// CHECK3-NEXT: call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32(i32 [[TMP1]]) #[[ATTR2]]
// CHECK3-NEXT: br label [[OMP_OFFLOAD_CONT]]
// CHECK3: omp_offload.cont:
// CHECK3-NEXT: ret i32 0
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32
// CHECK3-SAME: (i32 noundef [[T_VAR:%.*]]) #[[ATTR1]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[T_VAR_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT: store i32 [[T_VAR]], ptr [[T_VAR_ADDR]], align 4
// CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB3]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined, ptr [[T_VAR_ADDR]])
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined
// CHECK3-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[T_VAR:%.*]]) #[[ATTR1]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[T_VAR_ADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[T_VAR1:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[I:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x ptr], align 4
// CHECK3-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK3-NEXT: store ptr [[T_VAR]], ptr [[T_VAR_ADDR]], align 4
// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 4
// CHECK3-NEXT: store i32 0, ptr [[T_VAR1]], align 4
// CHECK3-NEXT: store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK3-NEXT: store i32 1, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK3-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
// CHECK3-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP2]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
// CHECK3-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1
// CHECK3-NEXT: br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK3: cond.true:
// CHECK3-NEXT: br label [[COND_END:%.*]]
// CHECK3: cond.false:
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: br label [[COND_END]]
// CHECK3: cond.end:
// CHECK3-NEXT: [[COND:%.*]] = phi i32 [ 1, [[COND_TRUE]] ], [ [[TMP4]], [[COND_FALSE]] ]
// CHECK3-NEXT: store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK3-NEXT: store i32 [[TMP5]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3: omp.inner.for.cond:
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK3-NEXT: [[CMP2:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK3-NEXT: br i1 [[CMP2]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK3: omp.inner.for.body:
// CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP8]], 1
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[I]], align 4
// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[T_VAR1]], align 4
// CHECK3-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP10]], [[TMP9]]
// CHECK3-NEXT: store i32 [[ADD3]], ptr [[T_VAR1]], align 4
// CHECK3-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK3: omp.body.continue:
// CHECK3-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK3: omp.inner.for.inc:
// CHECK3-NEXT: [[TMP11:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP11]], 1
// CHECK3-NEXT: store i32 [[ADD4]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK3: omp.inner.for.end:
// CHECK3-NEXT: br label [[OMP_LOOP_EXIT:%.*]]
// CHECK3: omp.loop.exit:
// CHECK3-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP2]])
// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i32 0, i32 0
// CHECK3-NEXT: store ptr [[T_VAR1]], ptr [[TMP12]], align 4
// CHECK3-NEXT: [[TMP13:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], i32 1, i32 4, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT: switch i32 [[TMP13]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK3-NEXT: i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK3-NEXT: i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK3-NEXT: ]
// CHECK3: .omp.reduction.case1:
// CHECK3-NEXT: [[TMP14:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK3-NEXT: [[TMP15:%.*]] = load i32, ptr [[T_VAR1]], align 4
// CHECK3-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP14]], [[TMP15]]
// CHECK3-NEXT: store i32 [[ADD5]], ptr [[TMP0]], align 4
// CHECK3-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], ptr @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3: .omp.reduction.case2:
// CHECK3-NEXT: [[TMP16:%.*]] = load i32, ptr [[T_VAR1]], align 4
// CHECK3-NEXT: [[TMP17:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP16]] monotonic, align 4
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3: .omp.reduction.default:
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z5tmainIiET_v_l32.omp_outlined.omp.reduction.reduction_func
// CHECK3-SAME: (ptr noundef [[TMP0:%.*]], ptr noundef [[TMP1:%.*]]) #[[ATTR3]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: [[DOTADDR1:%.*]] = alloca ptr, align 4
// CHECK3-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 4
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 4
// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i32 0, i32 0
// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP4]], align 4
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP2]], i32 0, i32 0
// CHECK3-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP8]], [[TMP9]]
// CHECK3-NEXT: store i32 [[ADD]], ptr [[TMP7]], align 4
// CHECK3-NEXT: ret void
//
//
// CHECK9-LABEL: define {{[^@]+}}@main
// CHECK9-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK9-NEXT: entry:
// CHECK9-NEXT: [[RETVAL:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[REF_TMP:%.*]] = alloca [[CLASS_ANON:%.*]], align 1
// CHECK9-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK9-NEXT: call void @"_ZZ4mainENK3$_0clEv"(ptr noundef nonnull align 1 dereferenceable(1) [[REF_TMP]])
// CHECK9-NEXT: ret i32 0
//
//
// CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l45
// CHECK9-SAME: (i64 noundef [[SIVAR:%.*]]) #[[ATTR2:[0-9]+]] {
// CHECK9-NEXT: entry:
// CHECK9-NEXT: [[SIVAR_ADDR:%.*]] = alloca i64, align 8
// CHECK9-NEXT: store i64 [[SIVAR]], ptr [[SIVAR_ADDR]], align 8
// CHECK9-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_teams(ptr @[[GLOB3:[0-9]+]], i32 1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l45.omp_outlined, ptr [[SIVAR_ADDR]])
// CHECK9-NEXT: ret void
//
//
// CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l45.omp_outlined
// CHECK9-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[SIVAR:%.*]]) #[[ATTR2]] {
// CHECK9-NEXT: entry:
// CHECK9-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK9-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK9-NEXT: [[SIVAR_ADDR:%.*]] = alloca ptr, align 8
// CHECK9-NEXT: [[SIVAR1:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[TMP:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[I:%.*]] = alloca i32, align 4
// CHECK9-NEXT: [[REF_TMP:%.*]] = alloca [[CLASS_ANON_0:%.*]], align 8
// CHECK9-NEXT: [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [1 x ptr], align 8
// CHECK9-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK9-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK9-NEXT: store ptr [[SIVAR]], ptr [[SIVAR_ADDR]], align 8
// CHECK9-NEXT: [[TMP0:%.*]] = load ptr, ptr [[SIVAR_ADDR]], align 8
// CHECK9-NEXT: store i32 0, ptr [[SIVAR1]], align 4
// CHECK9-NEXT: store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK9-NEXT: store i32 1, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK9-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK9-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK9-NEXT: [[TMP1:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK9-NEXT: [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
// CHECK9-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1:[0-9]+]], i32 [[TMP2]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
// CHECK9-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK9-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP3]], 1
// CHECK9-NEXT: br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK9: cond.true:
// CHECK9-NEXT: br label [[COND_END:%.*]]
// CHECK9: cond.false:
// CHECK9-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK9-NEXT: br label [[COND_END]]
// CHECK9: cond.end:
// CHECK9-NEXT: [[COND:%.*]] = phi i32 [ 1, [[COND_TRUE]] ], [ [[TMP4]], [[COND_FALSE]] ]
// CHECK9-NEXT: store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK9-NEXT: [[TMP5:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK9-NEXT: store i32 [[TMP5]], ptr [[DOTOMP_IV]], align 4
// CHECK9-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK9: omp.inner.for.cond:
// CHECK9-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK9-NEXT: [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK9-NEXT: [[CMP2:%.*]] = icmp sle i32 [[TMP6]], [[TMP7]]
// CHECK9-NEXT: br i1 [[CMP2]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK9: omp.inner.for.body:
// CHECK9-NEXT: [[TMP8:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK9-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP8]], 1
// CHECK9-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK9-NEXT: store i32 [[ADD]], ptr [[I]], align 4
// CHECK9-NEXT: [[TMP9:%.*]] = load i32, ptr [[I]], align 4
// CHECK9-NEXT: [[TMP10:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK9-NEXT: [[ADD3:%.*]] = add nsw i32 [[TMP10]], [[TMP9]]
// CHECK9-NEXT: store i32 [[ADD3]], ptr [[SIVAR1]], align 4
// CHECK9-NEXT: [[TMP11:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], ptr [[REF_TMP]], i32 0, i32 0
// CHECK9-NEXT: store ptr [[SIVAR1]], ptr [[TMP11]], align 8
// CHECK9-NEXT: call void @"_ZZZ4mainENK3$_0clEvENKUlvE_clEv"(ptr noundef nonnull align 8 dereferenceable(8) [[REF_TMP]])
// CHECK9-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK9: omp.body.continue:
// CHECK9-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK9: omp.inner.for.inc:
// CHECK9-NEXT: [[TMP12:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK9-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP12]], 1
// CHECK9-NEXT: store i32 [[ADD4]], ptr [[DOTOMP_IV]], align 4
// CHECK9-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK9: omp.inner.for.end:
// CHECK9-NEXT: br label [[OMP_LOOP_EXIT:%.*]]
// CHECK9: omp.loop.exit:
// CHECK9-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP2]])
// CHECK9-NEXT: [[TMP13:%.*]] = getelementptr inbounds [1 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
// CHECK9-NEXT: store ptr [[SIVAR1]], ptr [[TMP13]], align 8
// CHECK9-NEXT: [[TMP14:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP2]], i32 1, i64 8, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l45.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
// CHECK9-NEXT: switch i32 [[TMP14]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK9-NEXT: i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK9-NEXT: i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK9-NEXT: ]
// CHECK9: .omp.reduction.case1:
// CHECK9-NEXT: [[TMP15:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK9-NEXT: [[TMP16:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK9-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP15]], [[TMP16]]
// CHECK9-NEXT: store i32 [[ADD5]], ptr [[TMP0]], align 4
// CHECK9-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP2]], ptr @.gomp_critical_user_.reduction.var)
// CHECK9-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK9: .omp.reduction.case2:
// CHECK9-NEXT: [[TMP17:%.*]] = load i32, ptr [[SIVAR1]], align 4
// CHECK9-NEXT: [[TMP18:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP17]] monotonic, align 4
// CHECK9-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK9: .omp.reduction.default:
// CHECK9-NEXT: ret void
//
//
// CHECK9-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l45.omp_outlined.omp.reduction.reduction_func
// CHECK9-SAME: (ptr noundef [[TMP0:%.*]], ptr noundef [[TMP1:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK9-NEXT: entry:
// CHECK9-NEXT: [[DOTADDR:%.*]] = alloca ptr, align 8
// CHECK9-NEXT: [[DOTADDR1:%.*]] = alloca ptr, align 8
// CHECK9-NEXT: store ptr [[TMP0]], ptr [[DOTADDR]], align 8
// CHECK9-NEXT: store ptr [[TMP1]], ptr [[DOTADDR1]], align 8
// CHECK9-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTADDR]], align 8
// CHECK9-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTADDR1]], align 8
// CHECK9-NEXT: [[TMP4:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP3]], i64 0, i64 0
// CHECK9-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP4]], align 8
// CHECK9-NEXT: [[TMP6:%.*]] = getelementptr inbounds [1 x ptr], ptr [[TMP2]], i64 0, i64 0
// CHECK9-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK9-NEXT: [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
// CHECK9-NEXT: [[TMP9:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK9-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP8]], [[TMP9]]
// CHECK9-NEXT: store i32 [[ADD]], ptr [[TMP7]], align 4
// CHECK9-NEXT: ret void
//
Reference in New Issue View Git Blame Copy Permalink