[clang] [OpenMP] Avoid use of ConstantExpr::getSizeOf() (PR #222662)
Nikita Popov via cfe-commits
cfe-commits at lists.llvm.org
Thu Sep 10 06:55:37 PDT 2026
https://github.com/nikic created https://github.com/llvm/llvm-project/pull/222662
Support for this will be dropped as part of the ptradd migration. Replace with a constant size.
While I'm touching the code anyway, I also simplified the size calculation a bit. It's computing both the size in bytes and in elements. Avoid doing an unnecessary division followed by multiply for the size in bytes.
>From 21c28218a45ba891f77bef7e9194231849a4ff41 Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov at redhat.com>
Date: Thu, 10 Sep 2026 15:27:22 +0200
Subject: [PATCH] [OpenMP] Avoid use of ConstantExpr::getSizeOf()
Support for this will be dropped as part of the ptradd migration.
Replace with a constant size.
While I'm touching the code anyway, I also simplified the size
calculation a bit. It's computing both the size in bytes and in
elements. Avoid doing an unnecessary division followed by
multiply for the size in bytes.
---
clang/lib/CodeGen/CGOpenMPRuntime.cpp | 18 +-
...te_parallel_for_reduction_task_codegen.cpp | 305 +++++----
clang/test/OpenMP/for_reduction_codegen.cpp | 596 +++++++++---------
.../test/OpenMP/for_reduction_codegen_UDR.cpp | 503 ++++++++-------
.../OpenMP/for_reduction_task_codegen.cpp | 357 ++++++-----
.../master_taskloop_in_reduction_codegen.cpp | 109 ++--
.../master_taskloop_reduction_codegen.cpp | 9 +-
...ter_taskloop_simd_in_reduction_codegen.cpp | 131 ++--
...master_taskloop_simd_reduction_codegen.cpp | 9 +-
.../parallel_for_reduction_task_codegen.cpp | 347 +++++-----
...parallel_master_reduction_task_codegen.cpp | 325 +++++-----
...llel_master_taskloop_reduction_codegen.cpp | 9 +-
...master_taskloop_simd_reduction_codegen.cpp | 9 +-
.../OpenMP/parallel_reduction_codegen.cpp | 467 +++++++-------
.../parallel_reduction_task_codegen.cpp | 313 +++++----
...rallel_sections_reduction_task_codegen.cpp | 347 +++++-----
clang/test/OpenMP/reduction_implicit_map.cpp | 158 +++--
.../sections_reduction_task_codegen.cpp | 357 ++++++-----
.../OpenMP/target_in_reduction_codegen.cpp | 2 +-
...et_parallel_for_reduction_task_codegen.cpp | 305 +++++----
...target_parallel_reduction_task_codegen.cpp | 271 ++++----
...te_parallel_for_reduction_task_codegen.cpp | 555 ++++++++--------
.../test/OpenMP/task_in_reduction_codegen.cpp | 75 ++-
.../taskgroup_task_reduction_codegen.cpp | 2 +-
.../OpenMP/taskloop_in_reduction_codegen.cpp | 109 ++--
.../OpenMP/taskloop_reduction_codegen.cpp | 9 +-
.../taskloop_simd_in_reduction_codegen.cpp | 131 ++--
.../taskloop_simd_reduction_codegen.cpp | 9 +-
...te_parallel_for_reduction_task_codegen.cpp | 555 ++++++++--------
29 files changed, 3166 insertions(+), 3226 deletions(-)
diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
index 829a245ab14f2..80d20f3259822 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -817,20 +817,20 @@ void ReductionCodeGen::emitAggregateType(CodeGenFunction &CGF, unsigned N) {
llvm::Value *Size;
llvm::Value *SizeInChars;
auto *ElemType = OrigAddresses[N].first.getAddress().getElementType();
- auto *ElemSizeOf = llvm::ConstantExpr::getSizeOf(ElemType);
+ auto *ElemSizeOf = llvm::ConstantInt::get(
+ CGF.SizeTy, CGF.CGM.getDataLayout().getTypeAllocSize(ElemType));
if (AsArraySection) {
- Size = CGF.Builder.CreatePtrDiff(ElemType,
- OrigAddresses[N].second.getPointer(CGF),
- OrigAddresses[N].first.getPointer(CGF));
- Size = CGF.Builder.CreateZExtOrTrunc(Size, ElemSizeOf->getType());
- Size = CGF.Builder.CreateNUWAdd(
- Size, llvm::ConstantInt::get(Size->getType(), /*V=*/1));
- SizeInChars = CGF.Builder.CreateNUWMul(Size, ElemSizeOf);
+ SizeInChars =
+ CGF.Builder.CreatePtrDiff(OrigAddresses[N].second.getPointer(CGF),
+ OrigAddresses[N].first.getPointer(CGF));
+ SizeInChars = CGF.Builder.CreateNUWAdd(SizeInChars, ElemSizeOf);
} else {
SizeInChars =
CGF.getTypeSize(OrigAddresses[N].first.getType().getNonReferenceType());
- Size = CGF.Builder.CreateExactUDiv(SizeInChars, ElemSizeOf);
}
+ Size = ElemSizeOf->isOne()
+ ? SizeInChars
+ : CGF.Builder.CreateExactUDiv(SizeInChars, ElemSizeOf);
Sizes.emplace_back(SizeInChars, Size);
CodeGenFunction::OpaqueValueMapping OpaqueMap(
CGF,
diff --git a/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
index 6952dbc6176f3..8d0893970cf48 100644
--- a/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
@@ -191,223 +191,221 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP11:%.*]] = sub i64 [[TMP9]], [[TMP10]]
// CHECK1-NEXT: [[TMP12:%.*]] = add nuw i64 [[TMP11]], 1
-// CHECK1-NEXT: [[TMP13:%.*]] = mul nuw i64 [[TMP12]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP14:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP14]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP13:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP13]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP12]], align 16
// CHECK1-NEXT: store i64 [[TMP12]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP15]]
+// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP14]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP15]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP14]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP17:%.*]] = load ptr, ptr [[TMP16]], align 8
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[TMP17]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP20:%.*]] = sub i64 [[TMP18]], [[TMP19]]
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP20]]
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP16]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP19]]
// CHECK1-NEXT: store ptr [[_TMP6]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP21]], ptr [[_TMP6]], align 8
+// CHECK1-NEXT: store ptr [[TMP20]], ptr [[_TMP6]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP26]], align 8
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP28]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP27]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_7:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP30]], i64 0
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
-// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP31]], i64 0
-// CHECK1-NEXT: [[TMP32:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP33:%.*]] = sext i32 [[TMP32]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP33]]
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP34]], i64 9
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
-// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP35]], i64 [[LB_ADD_LEN10]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP29]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
-// CHECK1-NEXT: [[TMP40:%.*]] = add nuw i64 [[TMP39]], 1
-// CHECK1-NEXT: [[TMP41:%.*]] = mul nuw i64 [[TMP40]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP41]], ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP45]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP46]], align 8
-// CHECK1-NEXT: [[TMP47:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load i32, ptr [[TMP47]], align 4
-// CHECK1-NEXT: [[TMP49:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB2]], i32 [[TMP48]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP49]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[TMP50]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP51]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP52]], 9
+// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP29]], i64 0
+// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
+// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP30]], i64 0
+// CHECK1-NEXT: [[TMP31:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP32:%.*]] = sext i32 [[TMP31]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP32]]
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP33]], i64 9
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
+// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP34]], i64 [[LB_ADD_LEN10]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP28]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP35]], align 8
+// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = sub i64 [[TMP36]], [[TMP37]]
+// CHECK1-NEXT: [[TMP39:%.*]] = add nuw i64 [[TMP38]], 1
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
+// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
+// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB2]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP49]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
// 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: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP53]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP54]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP55]], [[TMP56]]
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP57:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP57]], 1
+// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP55]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP61]], ptr [[TMP60]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: [[TMP64:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB2]], i32 [[TMP63]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP64]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP65]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[TMP66]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP67]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP64]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP69:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP68]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP70]], ptr [[TMP69]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[TMP71]], align 4
-// CHECK1-NEXT: [[TMP73:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB2]], i32 [[TMP72]], ptr [[TMP64]])
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP59]], ptr [[TMP58]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
+// CHECK1-NEXT: [[TMP62:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB2]], i32 [[TMP61]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP62]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP63]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[TMP64]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP65]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP62]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP66]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP68]], ptr [[TMP67]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[TMP69]], align 4
+// CHECK1-NEXT: [[TMP71:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB2]], i32 [[TMP70]], ptr [[TMP62]])
// 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: [[TMP74:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP74]], 1
+// CHECK1-NEXT: [[TMP72:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP72]], 1
// CHECK1-NEXT: store i64 [[ADD14]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP73:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[TMP73]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP74]])
// CHECK1-NEXT: [[TMP75:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[TMP75]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP76]])
-// CHECK1-NEXT: [[TMP77:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[TMP77]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB2]], i32 [[TMP78]], i32 1)
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP80]], align 8
-// CHECK1-NEXT: [[TMP81:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP82:%.*]] = inttoptr i64 [[TMP12]] to ptr
-// CHECK1-NEXT: store ptr [[TMP82]], ptr [[TMP81]], align 8
-// CHECK1-NEXT: [[TMP83:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP83]], align 4
-// CHECK1-NEXT: [[TMP85:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4:[0-9]+]], i32 [[TMP84]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP85]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB2]], i32 [[TMP76]], i32 1)
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP78]], align 8
+// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP80:%.*]] = inttoptr i64 [[TMP12]] to ptr
+// CHECK1-NEXT: store ptr [[TMP80]], ptr [[TMP79]], align 8
+// CHECK1-NEXT: [[TMP81:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP81]], align 4
+// CHECK1-NEXT: [[TMP83:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4:[0-9]+]], i32 [[TMP82]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP83]], 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: [[TMP86:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP87:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP86]], [[TMP87]]
+// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP85:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP84]], [[TMP85]]
// CHECK1-NEXT: store i32 [[ADD15]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP88:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP88]]
+// CHECK1-NEXT: [[TMP86:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE22:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST16:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT20:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP89]] to i32
-// CHECK1-NEXT: [[TMP90:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP90]] to i32
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP87]] to i32
+// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP88]] to i32
// CHECK1-NEXT: [[ADD18:%.*]] = add nsw i32 [[CONV]], [[CONV17]]
// CHECK1-NEXT: [[CONV19:%.*]] = trunc i32 [[ADD18]] to i8
// CHECK1-NEXT: store i8 [[CONV19]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT20]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP88]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_DONE22]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done22:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP84]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP91:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP92:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP91]] monotonic, align 4
-// CHECK1-NEXT: [[TMP93:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP93]]
+// CHECK1-NEXT: [[TMP89:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP90:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP89]] monotonic, align 4
+// CHECK1-NEXT: [[TMP91:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY23]], label [[OMP_ARRAYCPY_DONE36:%.*]], label [[OMP_ARRAYCPY_BODY24:%.*]]
// CHECK1: omp.arraycpy.body24:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST25:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT34:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT33:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP92]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP95:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP100:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP95]], ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP96]] to i32
-// CHECK1-NEXT: [[TMP97:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP97]] to i32
+// CHECK1-NEXT: [[TMP93:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP98:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP93]], ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP95]] to i32
// CHECK1-NEXT: [[ADD31:%.*]] = add nsw i32 [[CONV29]], [[CONV30]]
// CHECK1-NEXT: [[CONV32:%.*]] = trunc i32 [[ADD31]] to i8
// CHECK1-NEXT: store i8 [[CONV32]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP98:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP99:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP95]], i8 [[TMP98]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP100]] = extractvalue { i8, i1 } [[TMP99]], 0
-// CHECK1-NEXT: [[TMP101:%.*]] = extractvalue { i8, i1 } [[TMP99]], 1
-// CHECK1-NEXT: br i1 [[TMP101]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP97:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP93]], i8 [[TMP96]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP98]] = extractvalue { i8, i1 } [[TMP97]], 0
+// CHECK1-NEXT: [[TMP99:%.*]] = extractvalue { i8, i1 } [[TMP97]], 1
+// CHECK1-NEXT: br i1 [[TMP99]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT34]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP93]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_DONE36]], label [[OMP_ARRAYCPY_BODY24]]
// CHECK1: omp.arraycpy.done36:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP102:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP102]])
+// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP100]])
// CHECK1-NEXT: ret void
//
//
@@ -568,19 +566,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/for_reduction_codegen.cpp b/clang/test/OpenMP/for_reduction_codegen.cpp
index 8d45693728c58..38289b4271fbb 100644
--- a/clang/test/OpenMP/for_reduction_codegen.cpp
+++ b/clang/test/OpenMP/for_reduction_codegen.cpp
@@ -716,7 +716,7 @@ int main() {
// CHECK1-NEXT: br i1 [[ARRAYCTOR_DONE8]], label [[ARRAYCTOR_CONT9:%.*]], label [[ARRAYCTOR_LOOP5]]
// CHECK1: arrayctor.cont9:
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[VAR3]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5:![0-9]+]], !align [[META6:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META4:![0-9]+]], !align [[META5:![0-9]+]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 6, ptr @main.omp_outlined, ptr [[T_VAR]], ptr [[TMP0]], ptr [[VAR1]], ptr [[T_VAR1]], ptr [[VEC]], ptr [[S_ARR]])
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 1
// CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
@@ -734,13 +734,13 @@ int main() {
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.6, ptr [[VAR2]])
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.7, ptr [[VAR2]])
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.8, ptr [[VVAR2]])
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.9, ptr [[TMP5]])
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.10, ptr [[TMP6]])
-// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.11, ptr [[TMP7]])
-// CHECK1-NEXT: [[TMP8:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP8:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 1, ptr @main.omp_outlined.12, ptr [[TMP8]])
// CHECK1-NEXT: call void @_Z21reductionArrayElementv()
// CHECK1-NEXT: [[CALL10:%.*]] = call noundef i32 @_Z5tmainIiLi42EET_v()
@@ -844,21 +844,21 @@ int main() {
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VEC]], ptr [[VEC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_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: store float 0.000000e+00, ptr [[T_VAR3]], align 4
-// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR4]])
// CHECK1-NEXT: store ptr [[VAR4]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR16]])
@@ -898,7 +898,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP17]] to i64
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP4]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4
-// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[_TMP5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[_TMP5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP19:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM9:%.*]] = sext i32 [[TMP19]] to i64
// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds [4 x [[STRUCT_S]]], ptr [[TMP5]], i64 0, i64 [[IDXPROM9]]
@@ -1141,9 +1141,9 @@ int main() {
// CHECK1-NEXT: store ptr [[ARRS]], ptr [[ARRS_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR2]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -1161,147 +1161,145 @@ int main() {
// CHECK1-NEXT: [[TMP9:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
// CHECK1-NEXT: [[TMP11:%.*]] = sub i64 [[TMP9]], [[TMP10]]
-// CHECK1-NEXT: [[TMP12:%.*]] = sdiv exact i64 [[TMP11]], 4
-// CHECK1-NEXT: [[TMP13:%.*]] = add nuw i64 [[TMP12]], 1
-// CHECK1-NEXT: [[TMP14:%.*]] = mul nuw i64 [[TMP13]], ptrtoint (ptr getelementptr (i32, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP15:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP15]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP12:%.*]] = add nuw i64 [[TMP11]], 4
+// CHECK1-NEXT: [[TMP13:%.*]] = udiv exact i64 [[TMP12]], 4
+// CHECK1-NEXT: [[TMP14:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP14]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA7:%.*]] = alloca i32, i64 [[TMP13]], align 16
// CHECK1-NEXT: store i64 [[TMP13]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP13]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA7]], [[TMP16]]
+// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA7]], [[TMP15]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA7]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i32 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 4
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP16]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP15]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
-// CHECK1-NEXT: [[TMP20:%.*]] = sdiv exact i64 [[TMP19]], 4
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP20]]
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = sdiv exact i64 [[TMP18]], 4
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP19]]
// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw [10 x [4 x [[STRUCT_S:%.*]]]], ptr [[TMP4]], i64 0, i64 1
// CHECK1-NEXT: [[ARRAYDECAY:%.*]] = getelementptr inbounds [4 x [[STRUCT_S]]], ptr [[ARRAYIDX8]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw [[STRUCT_S]], ptr [[ARRAYDECAY]], i64 1
// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP3]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[ARRAYIDX10]], align 4
-// CHECK1-NEXT: [[TMP23:%.*]] = sext i32 [[TMP22]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN11:%.*]] = add nsw i64 0, [[TMP23]]
+// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[ARRAYIDX10]], align 4
+// CHECK1-NEXT: [[TMP22:%.*]] = sext i32 [[TMP21]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN11:%.*]] = add nsw i64 0, [[TMP22]]
// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw [10 x [4 x [[STRUCT_S]]]], ptr [[TMP4]], i64 0, i64 [[LB_ADD_LEN11]]
// CHECK1-NEXT: [[ARRAYDECAY13:%.*]] = getelementptr inbounds [4 x [[STRUCT_S]]], ptr [[ARRAYIDX12]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYIDX14:%.*]] = getelementptr inbounds nuw [[STRUCT_S]], ptr [[ARRAYDECAY13]], i64 2
-// CHECK1-NEXT: [[TMP24:%.*]] = ptrtoaddr ptr [[ARRAYIDX14]] to i64
-// CHECK1-NEXT: [[TMP25:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP26:%.*]] = sub i64 [[TMP24]], [[TMP25]]
-// CHECK1-NEXT: [[TMP27:%.*]] = sdiv exact i64 [[TMP26]], 4
-// CHECK1-NEXT: [[TMP28:%.*]] = add nuw i64 [[TMP27]], 1
-// CHECK1-NEXT: [[TMP29:%.*]] = mul nuw i64 [[TMP28]], ptrtoint (ptr getelementptr ([[STRUCT_S]], ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[VLA15:%.*]] = alloca [[STRUCT_S]], i64 [[TMP28]], align 16
-// CHECK1-NEXT: store i64 [[TMP28]], ptr [[__VLA_EXPR1]], align 8
-// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA15]], i64 [[TMP28]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY16:%.*]] = icmp eq ptr [[VLA15]], [[TMP30]]
+// CHECK1-NEXT: [[TMP23:%.*]] = ptrtoaddr ptr [[ARRAYIDX14]] to i64
+// CHECK1-NEXT: [[TMP24:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP25:%.*]] = sub i64 [[TMP23]], [[TMP24]]
+// CHECK1-NEXT: [[TMP26:%.*]] = add nuw i64 [[TMP25]], 4
+// CHECK1-NEXT: [[TMP27:%.*]] = udiv exact i64 [[TMP26]], 4
+// CHECK1-NEXT: [[VLA15:%.*]] = alloca [[STRUCT_S]], i64 [[TMP27]], align 16
+// CHECK1-NEXT: store i64 [[TMP27]], ptr [[__VLA_EXPR1]], align 8
+// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA15]], i64 [[TMP27]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY16:%.*]] = icmp eq ptr [[VLA15]], [[TMP28]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY16]], label [[OMP_ARRAYINIT_DONE21:%.*]], label [[OMP_ARRAYINIT_BODY17:%.*]]
// CHECK1: omp.arrayinit.body17:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST18:%.*]] = phi ptr [ [[VLA15]], [[OMP_ARRAYINIT_DONE]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT19:%.*]], [[OMP_ARRAYINIT_BODY17]] ]
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST18]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT19]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST18]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT19]], [[TMP30]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT19]], [[TMP28]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYINIT_DONE21]], label [[OMP_ARRAYINIT_BODY17]]
// CHECK1: omp.arrayinit.done21:
-// CHECK1-NEXT: [[TMP31:%.*]] = ptrtoaddr ptr [[TMP4]] to i64
-// CHECK1-NEXT: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP33:%.*]] = sub i64 [[TMP31]], [[TMP32]]
-// CHECK1-NEXT: [[TMP34:%.*]] = sdiv exact i64 [[TMP33]], 4
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA15]], i64 [[TMP34]]
-// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[TMP36]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP37]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP38]], 9
+// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP4]] to i64
+// CHECK1-NEXT: [[TMP30:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP31:%.*]] = sub i64 [[TMP29]], [[TMP30]]
+// CHECK1-NEXT: [[TMP32:%.*]] = sdiv exact i64 [[TMP31]], 4
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA15]], i64 [[TMP32]]
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = load i32, ptr [[TMP34]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP35]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP36:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP36]], 9
// 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: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP39]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP37]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
-// CHECK1-NEXT: store i32 [[TMP40]], ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
+// CHECK1-NEXT: store i32 [[TMP38]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP22:%.*]] = icmp sle i32 [[TMP41]], [[TMP42]]
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP22:%.*]] = icmp sle i32 [[TMP39]], [[TMP40]]
// CHECK1-NEXT: br i1 [[CMP22]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP43]], 1
+// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP41]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], align 4
-// CHECK1-NEXT: [[TMP44:%.*]] = mul nsw i64 1, [[TMP1]]
-// CHECK1-NEXT: [[ARRAYIDX23:%.*]] = getelementptr inbounds i32, ptr [[TMP21]], i64 [[TMP44]]
-// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[I]], align 4
-// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP45]] to i64
+// CHECK1-NEXT: [[TMP42:%.*]] = mul nsw i64 1, [[TMP1]]
+// CHECK1-NEXT: [[ARRAYIDX23:%.*]] = getelementptr inbounds i32, ptr [[TMP20]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[I]], align 4
+// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP43]] to i64
// CHECK1-NEXT: [[ARRAYIDX24:%.*]] = getelementptr inbounds i32, ptr [[ARRAYIDX23]], i64 [[IDXPROM]]
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[ARRAYIDX24]], align 4
-// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP46]], 1
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[ARRAYIDX24]], align 4
+// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP44]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[ARRAYIDX24]], 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: [[TMP47:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[ADD25:%.*]] = add nsw i32 [[TMP47]], 1
+// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[ADD25:%.*]] = add nsw i32 [[TMP45]], 1
// CHECK1-NEXT: store i32 [[ADD25]], 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: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP49]])
-// CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA7]], ptr [[TMP50]], align 8
-// CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP52:%.*]] = inttoptr i64 [[TMP13]] to ptr
-// CHECK1-NEXT: store ptr [[TMP52]], ptr [[TMP51]], align 8
-// CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: store ptr [[VLA15]], ptr [[TMP53]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
-// CHECK1-NEXT: [[TMP55:%.*]] = inttoptr i64 [[TMP28]] to ptr
-// CHECK1-NEXT: store ptr [[TMP55]], ptr [[TMP54]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = load i32, ptr [[TMP56]], align 4
-// CHECK1-NEXT: [[TMP58:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP57]], i32 2, i64 32, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.1.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP58]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP47]])
+// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA7]], ptr [[TMP48]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP50:%.*]] = inttoptr i64 [[TMP13]] to ptr
+// CHECK1-NEXT: store ptr [[TMP50]], ptr [[TMP49]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: store ptr [[VLA15]], ptr [[TMP51]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
+// CHECK1-NEXT: [[TMP53:%.*]] = inttoptr i64 [[TMP27]] to ptr
+// CHECK1-NEXT: store ptr [[TMP53]], ptr [[TMP52]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = load i32, ptr [[TMP54]], align 4
+// CHECK1-NEXT: [[TMP56:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP55]], i32 2, i64 32, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.1.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP56]], 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: [[TMP59:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP59]]
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP57]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE30:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA7]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX3]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT28:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP60:%.*]] = load i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], align 4
-// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 4
-// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[TMP60]], [[TMP61]]
+// CHECK1-NEXT: [[TMP58:%.*]] = load i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], align 4
+// CHECK1-NEXT: [[TMP59:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 4
+// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[TMP58]], [[TMP59]]
// CHECK1-NEXT: store i32 [[ADD27]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], align 4
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT28]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE29:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT28]], [[TMP59]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE29:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT28]], [[TMP57]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE29]], label [[OMP_ARRAYCPY_DONE30]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done30:
-// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX9]], i64 [[TMP28]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY31:%.*]] = icmp eq ptr [[ARRAYIDX9]], [[TMP62]]
+// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX9]], i64 [[TMP27]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY31:%.*]] = icmp eq ptr [[ARRAYIDX9]], [[TMP60]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY31]], label [[OMP_ARRAYCPY_DONE38:%.*]], label [[OMP_ARRAYCPY_BODY32:%.*]]
// CHECK1: omp.arraycpy.body32:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST33:%.*]] = phi ptr [ [[VLA15]], [[OMP_ARRAYCPY_DONE30]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT36:%.*]], [[OMP_ARRAYCPY_BODY32]] ]
@@ -1310,56 +1308,56 @@ int main() {
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[OMP_ARRAYCPY_DESTELEMENTPAST34]], ptr align 4 [[CALL]], i64 4, i1 false)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT35]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST34]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT36]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST33]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE37:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT35]], [[TMP62]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE37:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT35]], [[TMP60]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE37]], label [[OMP_ARRAYCPY_DONE38]], label [[OMP_ARRAYCPY_BODY32]]
// CHECK1: omp.arraycpy.done38:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP57]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP55]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY39:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP63]]
+// CHECK1-NEXT: [[TMP61:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY39:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP61]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY39]], label [[OMP_ARRAYCPY_DONE46:%.*]], label [[OMP_ARRAYCPY_BODY40:%.*]]
// CHECK1: omp.arraycpy.body40:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST41:%.*]] = phi ptr [ [[VLA7]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT44:%.*]], [[OMP_ARRAYCPY_BODY40]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST42:%.*]] = phi ptr [ [[ARRAYIDX3]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT43:%.*]], [[OMP_ARRAYCPY_BODY40]] ]
-// CHECK1-NEXT: [[TMP64:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST41]], align 4
-// CHECK1-NEXT: [[TMP65:%.*]] = atomicrmw add ptr [[OMP_ARRAYCPY_DESTELEMENTPAST42]], i32 [[TMP64]] monotonic, align 4
+// CHECK1-NEXT: [[TMP62:%.*]] = load i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST41]], align 4
+// CHECK1-NEXT: [[TMP63:%.*]] = atomicrmw add ptr [[OMP_ARRAYCPY_DESTELEMENTPAST42]], i32 [[TMP62]] monotonic, align 4
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT43]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST42]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT44]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST41]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE45:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT43]], [[TMP63]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE45:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT43]], [[TMP61]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE45]], label [[OMP_ARRAYCPY_DONE46]], label [[OMP_ARRAYCPY_BODY40]]
// CHECK1: omp.arraycpy.done46:
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX9]], i64 [[TMP28]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY47:%.*]] = icmp eq ptr [[ARRAYIDX9]], [[TMP66]]
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX9]], i64 [[TMP27]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY47:%.*]] = icmp eq ptr [[ARRAYIDX9]], [[TMP64]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY47]], label [[OMP_ARRAYCPY_DONE55:%.*]], label [[OMP_ARRAYCPY_BODY48:%.*]]
// CHECK1: omp.arraycpy.body48:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST49:%.*]] = phi ptr [ [[VLA15]], [[OMP_ARRAYCPY_DONE46]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT53:%.*]], [[OMP_ARRAYCPY_BODY48]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST50:%.*]] = phi ptr [ [[ARRAYIDX9]], [[OMP_ARRAYCPY_DONE46]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT52:%.*]], [[OMP_ARRAYCPY_BODY48]] ]
-// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP68:%.*]] = load i32, ptr [[TMP67]], align 4
-// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB4]], i32 [[TMP68]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP66:%.*]] = load i32, ptr [[TMP65]], align 4
+// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB4]], i32 [[TMP66]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[CALL51:%.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @_ZN1SIfEanERKS0_(ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST50]], ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_SRCELEMENTPAST49]])
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[OMP_ARRAYCPY_DESTELEMENTPAST50]], ptr align 4 [[CALL51]], i64 4, i1 false)
-// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB4]], i32 [[TMP68]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB4]], i32 [[TMP66]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT52]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST50]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT53]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST49]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE54:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT52]], [[TMP66]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE54:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT52]], [[TMP64]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE54]], label [[OMP_ARRAYCPY_DONE55]], label [[OMP_ARRAYCPY_BODY48]]
// CHECK1: omp.arraycpy.done55:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP69:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[VLA15]], i64 [[TMP28]]
-// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA15]], [[TMP69]]
+// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[VLA15]], i64 [[TMP27]]
+// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA15]], [[TMP67]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE56:%.*]], label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1: arraydestroy.body:
-// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP69]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
+// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP67]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT: call void @_ZN1SIfED1Ev(ptr noundef nonnull align 4 dead_on_return(4) dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR2]]
// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[VLA15]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE56]], label [[ARRAYDESTROY_BODY]]
// CHECK1: arraydestroy.done56:
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP70]])
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP68]])
// CHECK1-NEXT: ret void
//
//
@@ -1445,15 +1443,15 @@ int main() {
// CHECK1-NEXT: store ptr [[ARRS]], ptr [[ARRS_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR2]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP4:%.*]] = mul nuw i64 [[TMP0]], [[TMP1]]
// CHECK1-NEXT: [[TMP5:%.*]] = mul nuw i64 [[TMP4]], 4
-// CHECK1-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], ptrtoint (ptr getelementptr (i32, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], 4
// CHECK1-NEXT: [[TMP7:%.*]] = call ptr @llvm.stacksave.p0()
// CHECK1-NEXT: store ptr [[TMP7]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA3:%.*]] = alloca i32, i64 [[TMP6]], align 16
@@ -1702,7 +1700,7 @@ int main() {
// CHECK1-NEXT: store ptr [[ARR]], ptr [[ARR_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR2]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -1880,7 +1878,7 @@ int main() {
// 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 [[VAR2]], ptr [[VAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META5]], !align [[META7:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META4]], !align [[META6:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -1896,90 +1894,89 @@ int main() {
// CHECK1-NEXT: [[TMP5:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
// CHECK1-NEXT: [[TMP6:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
// CHECK1-NEXT: [[TMP7:%.*]] = sub i64 [[TMP5]], [[TMP6]]
-// CHECK1-NEXT: [[TMP8:%.*]] = sdiv exact i64 [[TMP7]], 4
-// CHECK1-NEXT: [[TMP9:%.*]] = add nuw i64 [[TMP8]], 1
-// CHECK1-NEXT: [[TMP10:%.*]] = mul nuw i64 [[TMP9]], ptrtoint (ptr getelementptr ([[STRUCT_S]], ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP8:%.*]] = add nuw i64 [[TMP7]], 4
+// CHECK1-NEXT: [[TMP9:%.*]] = udiv exact i64 [[TMP8]], 4
+// CHECK1-NEXT: [[TMP10:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP10]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca [[STRUCT_S]], i64 [[TMP9]], align 16
// CHECK1-NEXT: store i64 [[TMP9]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP9]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
+// CHECK1-NEXT: [[TMP11:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP11]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP11]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[TMP0]], align 8
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
-// CHECK1-NEXT: [[TMP18:%.*]] = sdiv exact i64 [[TMP17]], 4
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[TMP0]], align 8
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = ptrtoaddr ptr [[TMP13]] to i64
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = sub i64 [[TMP14]], [[TMP15]]
+// CHECK1-NEXT: [[TMP17:%.*]] = sdiv exact i64 [[TMP16]], 4
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[_TMP4]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[TMP20]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP21]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP22]], 9
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[TMP19]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP20]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP21]], 9
// 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: [[TMP23:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP23]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP22]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[TMP24:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
-// CHECK1-NEXT: store i32 [[TMP24]], ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
+// CHECK1-NEXT: store i32 [[TMP23]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP25:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[TMP26:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP6:%.*]] = icmp sle i32 [[TMP25]], [[TMP26]]
+// CHECK1-NEXT: [[TMP24:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP25:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP6:%.*]] = icmp sle i32 [[TMP24]], [[TMP25]]
// CHECK1-NEXT: br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP27:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP27]], 1
+// CHECK1-NEXT: [[TMP26:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP26]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], 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: [[TMP28:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[ADD7:%.*]] = add nsw i32 [[TMP28]], 1
+// CHECK1-NEXT: [[TMP27:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[ADD7:%.*]] = add nsw i32 [[TMP27]], 1
// CHECK1-NEXT: store i32 [[ADD7]], 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: [[TMP29:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP29]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP30]])
-// CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP31]], align 8
-// CHECK1-NEXT: [[TMP32:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP33:%.*]] = inttoptr i64 [[TMP9]] to ptr
-// CHECK1-NEXT: store ptr [[TMP33]], ptr [[TMP32]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = load i32, ptr [[TMP34]], align 4
-// CHECK1-NEXT: [[TMP36:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB2]], i32 [[TMP35]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.4.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP36]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP28]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP29]])
+// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP30]], align 8
+// CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP32:%.*]] = inttoptr i64 [[TMP9]] to ptr
+// CHECK1-NEXT: store ptr [[TMP32]], ptr [[TMP31]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load i32, ptr [[TMP33]], align 4
+// CHECK1-NEXT: [[TMP35:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB2]], i32 [[TMP34]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.4.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP35]], 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: [[TMP37:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP37]]
+// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP36]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE11:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
@@ -1988,47 +1985,47 @@ int main() {
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[OMP_ARRAYCPY_DESTELEMENTPAST8]], ptr align 4 [[CALL]], i64 4, i1 false)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT9]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST8]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE10:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT9]], [[TMP37]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE10:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT9]], [[TMP36]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE10]], label [[OMP_ARRAYCPY_DONE11]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done11:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP35]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP34]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY12:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP38]]
+// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY12:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP37]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY12]], label [[OMP_ARRAYCPY_DONE20:%.*]], label [[OMP_ARRAYCPY_BODY13:%.*]]
// CHECK1: omp.arraycpy.body13:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST14:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT18:%.*]], [[OMP_ARRAYCPY_BODY13]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST15:%.*]] = phi ptr [ [[ARRAYIDX1]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT17:%.*]], [[OMP_ARRAYCPY_BODY13]] ]
-// CHECK1-NEXT: [[TMP39:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP39]], align 4
-// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB4]], i32 [[TMP40]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[TMP38]], align 4
+// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB4]], i32 [[TMP39]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[CALL16:%.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @_ZN1SIfEanERKS0_(ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST15]], ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_SRCELEMENTPAST14]])
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[OMP_ARRAYCPY_DESTELEMENTPAST15]], ptr align 4 [[CALL16]], i64 4, i1 false)
-// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB4]], i32 [[TMP40]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB4]], i32 [[TMP39]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT17]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT18]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST14]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE19:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT17]], [[TMP38]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE19:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT17]], [[TMP37]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE19]], label [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYCPY_BODY13]]
// CHECK1: omp.arraycpy.done20:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP35]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP34]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[VLA]], i64 [[TMP9]]
-// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP41]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[VLA]], i64 [[TMP9]]
+// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP40]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE21:%.*]], label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1: arraydestroy.body:
-// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP41]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
+// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP40]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT: call void @_ZN1SIfED1Ev(ptr noundef nonnull align 4 dead_on_return(4) dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR2]]
// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[VLA]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE21]], label [[ARRAYDESTROY_BODY]]
// CHECK1: arraydestroy.done21:
-// CHECK1-NEXT: [[TMP42:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP42]])
-// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
-// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB3]], i32 [[TMP44]])
+// CHECK1-NEXT: [[TMP41:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP41]])
+// CHECK1-NEXT: [[TMP42:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[TMP42]], align 4
+// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB3]], i32 [[TMP43]])
// CHECK1-NEXT: ret void
//
//
@@ -2084,7 +2081,7 @@ int main() {
// 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 [[VAR2]], ptr [[VAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META5]], !align [[META7]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META4]], !align [[META6]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -2271,7 +2268,7 @@ int main() {
// 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 [[VAR2]], ptr [[VAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META5]], !align [[META7]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META4]], !align [[META6]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -2458,7 +2455,7 @@ int main() {
// 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 [[VAR2]], ptr [[VAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META5]], !align [[META7]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META4]], !align [[META6]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -2583,7 +2580,7 @@ int main() {
// 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 [[VVAR2]], ptr [[VVAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VVAR2_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VVAR2_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -2761,17 +2758,17 @@ int main() {
// 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 [[VAR3]], ptr [[VAR3_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S]]], ptr [[TMP2]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S]]], ptr [[TMP3]], i64 0, i64 2
// CHECK1-NEXT: [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[VAR34]], i32 0, i32 0
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAY_BEGIN]], i64 2
@@ -2784,7 +2781,7 @@ int main() {
// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP4]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP6:%.*]] = ptrtoaddr ptr [[TMP5]] to i64
// CHECK1-NEXT: [[TMP7:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK1-NEXT: [[TMP8:%.*]] = sub i64 [[TMP6]], [[TMP7]]
@@ -2946,17 +2943,17 @@ int main() {
// 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 [[VAR3]], ptr [[VAR3_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S]]], ptr [[TMP2]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S]]], ptr [[TMP3]], i64 0, i64 1
// CHECK1-NEXT: [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[VAR34]], i32 0, i32 0
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAY_BEGIN]], i64 2
@@ -2969,7 +2966,7 @@ int main() {
// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP4]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP6:%.*]] = ptrtoaddr ptr [[TMP5]] to i64
// CHECK1-NEXT: [[TMP7:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK1-NEXT: [[TMP8:%.*]] = sub i64 [[TMP6]], [[TMP7]]
@@ -3132,103 +3129,102 @@ int main() {
// 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 [[VAR3]], ptr [[VAR3_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S:%.*]]], ptr [[TMP2]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S]]], ptr [[TMP3]], i64 0, i64 3
// CHECK1-NEXT: [[TMP4:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
// CHECK1-NEXT: [[TMP5:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK1-NEXT: [[TMP6:%.*]] = sub i64 [[TMP4]], [[TMP5]]
-// CHECK1-NEXT: [[TMP7:%.*]] = sdiv exact i64 [[TMP6]], 4
-// CHECK1-NEXT: [[TMP8:%.*]] = add nuw i64 [[TMP7]], 1
-// CHECK1-NEXT: [[TMP9:%.*]] = mul nuw i64 [[TMP8]], ptrtoint (ptr getelementptr ([[STRUCT_S]], ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP10:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP10]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP7:%.*]] = add nuw i64 [[TMP6]], 4
+// CHECK1-NEXT: [[TMP8:%.*]] = udiv exact i64 [[TMP7]], 4
+// CHECK1-NEXT: [[TMP9:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP9]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca [[STRUCT_S]], i64 [[TMP8]], align 16
// CHECK1-NEXT: store i64 [[TMP8]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP11:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP8]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP11]]
+// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP8]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP10]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP10]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP13:%.*]] = ptrtoaddr ptr [[TMP12]] to i64
-// CHECK1-NEXT: [[TMP14:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
-// CHECK1-NEXT: [[TMP15:%.*]] = sub i64 [[TMP13]], [[TMP14]]
-// CHECK1-NEXT: [[TMP16:%.*]] = sdiv exact i64 [[TMP15]], 4
-// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP16]]
-// CHECK1-NEXT: store ptr [[TMP17]], ptr [[_TMP4]], align 8
-// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP19:%.*]] = load i32, ptr [[TMP18]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP19]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP20]], 9
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP12:%.*]] = ptrtoaddr ptr [[TMP11]] to i64
+// CHECK1-NEXT: [[TMP13:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
+// CHECK1-NEXT: [[TMP14:%.*]] = sub i64 [[TMP12]], [[TMP13]]
+// CHECK1-NEXT: [[TMP15:%.*]] = sdiv exact i64 [[TMP14]], 4
+// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr [[STRUCT_S]], ptr [[VLA]], i64 [[TMP15]]
+// CHECK1-NEXT: store ptr [[TMP16]], ptr [[_TMP4]], align 8
+// CHECK1-NEXT: [[TMP17:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP18:%.*]] = load i32, ptr [[TMP17]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB1]], i32 [[TMP18]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP19:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP19]], 9
// 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: [[TMP21:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP21]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP20]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
-// CHECK1-NEXT: store i32 [[TMP22]], ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
+// CHECK1-NEXT: store i32 [[TMP21]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[TMP24:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP5:%.*]] = icmp sle i32 [[TMP23]], [[TMP24]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP5:%.*]] = icmp sle i32 [[TMP22]], [[TMP23]]
// CHECK1-NEXT: br i1 [[CMP5]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP25:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP25]], 1
+// CHECK1-NEXT: [[TMP24:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP24]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], 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: [[TMP26:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[ADD6:%.*]] = add nsw i32 [[TMP26]], 1
+// CHECK1-NEXT: [[TMP25:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[ADD6:%.*]] = add nsw i32 [[TMP25]], 1
// CHECK1-NEXT: store i32 [[ADD6]], 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: [[TMP27:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[TMP27]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP28]])
-// CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP29]], align 8
-// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP31:%.*]] = inttoptr i64 [[TMP8]] to ptr
-// CHECK1-NEXT: store ptr [[TMP31]], ptr [[TMP30]], align 8
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP33:%.*]] = load i32, ptr [[TMP32]], align 4
-// CHECK1-NEXT: [[TMP34:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB2]], i32 [[TMP33]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.11.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP34]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP27:%.*]] = load i32, ptr [[TMP26]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB1]], i32 [[TMP27]])
+// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP28]], align 8
+// CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP30:%.*]] = inttoptr i64 [[TMP8]] to ptr
+// CHECK1-NEXT: store ptr [[TMP30]], ptr [[TMP29]], align 8
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP32:%.*]] = load i32, ptr [[TMP31]], align 4
+// CHECK1-NEXT: [[TMP33:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB2]], i32 [[TMP32]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.11.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP33]], 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: [[TMP35:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX]], i64 [[TMP8]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP35]]
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX]], i64 [[TMP8]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP34]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE10:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
@@ -3237,47 +3233,47 @@ int main() {
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[OMP_ARRAYCPY_DESTELEMENTPAST7]], ptr align 4 [[CALL]], i64 4, i1 false)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT8]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST7]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE9:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT8]], [[TMP35]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE9:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT8]], [[TMP34]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE9]], label [[OMP_ARRAYCPY_DONE10]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done10:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP33]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP32]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX]], i64 [[TMP8]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY11:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP36]]
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAYIDX]], i64 [[TMP8]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY11:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP35]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY11]], label [[OMP_ARRAYCPY_DONE19:%.*]], label [[OMP_ARRAYCPY_BODY12:%.*]]
// CHECK1: omp.arraycpy.body12:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST13:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT17:%.*]], [[OMP_ARRAYCPY_BODY12]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST14:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT16:%.*]], [[OMP_ARRAYCPY_BODY12]] ]
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[TMP37]], align 4
-// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB4]], i32 [[TMP38]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[TMP36]], align 4
+// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB4]], i32 [[TMP37]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[CALL15:%.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @_ZN1SIfEanERKS0_(ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_DESTELEMENTPAST14]], ptr noundef nonnull align 4 dereferenceable(4) [[OMP_ARRAYCPY_SRCELEMENTPAST13]])
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[OMP_ARRAYCPY_DESTELEMENTPAST14]], ptr align 4 [[CALL15]], i64 4, i1 false)
-// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB4]], i32 [[TMP38]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB4]], i32 [[TMP37]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT16]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST14]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT17]] = getelementptr [[STRUCT_S]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST13]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE18:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP36]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE18:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP35]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_DONE19]], label [[OMP_ARRAYCPY_BODY12]]
// CHECK1: omp.arraycpy.done19:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP33]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB2]], i32 [[TMP32]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[VLA]], i64 [[TMP8]]
-// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP39]]
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[VLA]], i64 [[TMP8]]
+// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP38]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE20:%.*]], label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1: arraydestroy.body:
-// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP39]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
+// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP38]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT: call void @_ZN1SIfED1Ev(ptr noundef nonnull align 4 dead_on_return(4) dereferenceable(4) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR2]]
// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[VLA]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE20]], label [[ARRAYDESTROY_BODY]]
// CHECK1: arraydestroy.done20:
-// CHECK1-NEXT: [[TMP40:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP40]])
-// CHECK1-NEXT: [[TMP41:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[TMP41]], align 4
-// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB3]], i32 [[TMP42]])
+// CHECK1-NEXT: [[TMP39:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP39]])
+// CHECK1-NEXT: [[TMP40:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[TMP40]], align 4
+// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB3]], i32 [[TMP41]])
// CHECK1-NEXT: ret void
//
//
@@ -3334,15 +3330,15 @@ int main() {
// 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 [[VAR3]], ptr [[VAR3_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [4 x [[STRUCT_S]]], ptr [[VAR33]], i32 0, i32 0
// CHECK1-NEXT: [[TMP3:%.*]] = getelementptr [[STRUCT_S]], ptr [[ARRAY_BEGIN]], i64 4
// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[ARRAY_BEGIN]], [[TMP3]]
@@ -3520,11 +3516,11 @@ int main() {
// CHECK1-NEXT: [[ARRAYCTOR_DONE:%.*]] = icmp eq ptr [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK1-NEXT: br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK1: arrayctor.cont:
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 6, ptr @_Z5tmainIiLi42EET_v.omp_outlined, ptr [[T_VAR]], ptr [[TMP0]], ptr [[VAR1]], ptr [[T_VAR1]], ptr [[VEC]], ptr [[S_ARR]])
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 4, ptr @_Z5tmainIiLi42EET_v.omp_outlined.13, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[TMP1]])
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB4]], i32 5, ptr @_Z5tmainIiLi42EET_v.omp_outlined.14, ptr [[ARR]], ptr [[VEC]], ptr [[T_VAR]], ptr [[S_ARR]], ptr [[TMP2]])
// CHECK1-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK1-NEXT: [[ARRAY_BEGIN1:%.*]] = getelementptr inbounds [42 x [[STRUCT_S_0]]], ptr [[ARR]], i32 0, i32 0
@@ -3652,21 +3648,21 @@ int main() {
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VEC]], ptr [[VEC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_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: store i32 0, ptr [[T_VAR3]], align 4
-// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR4]])
// CHECK1-NEXT: store ptr [[VAR4]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR16]])
@@ -3705,7 +3701,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP17]] to i64
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP4]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[TMP16]], ptr [[ARRAYIDX]], align 4
-// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[_TMP5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[_TMP5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP19:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM9:%.*]] = sext i32 [[TMP19]] to i64
// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds [2 x [[STRUCT_S_0]]], ptr [[TMP5]], i64 0, i64 [[IDXPROM9]]
@@ -3920,12 +3916,12 @@ int main() {
// CHECK1-NEXT: store ptr [[VEC]], ptr [[VEC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP3]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP4]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_UB]], align 4
@@ -3964,7 +3960,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP14]] to i64
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP1]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[TMP13]], ptr [[ARRAYIDX]], align 4
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM5:%.*]] = sext i32 [[TMP16]] to i64
// CHECK1-NEXT: [[ARRAYIDX6:%.*]] = getelementptr inbounds [2 x [[STRUCT_S_0:%.*]]], ptr [[TMP2]], i64 0, i64 [[IDXPROM5]]
@@ -4091,13 +4087,13 @@ int main() {
// CHECK1-NEXT: store ptr [[T_VAR]], ptr [[T_VAR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP4]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP5]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_UB]], align 4
@@ -4155,7 +4151,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP21]] to i64
// CHECK1-NEXT: [[ARRAYIDX6:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP1]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[TMP20]], ptr [[ARRAYIDX6]], align 4
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM7:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds [2 x [[STRUCT_S_0]]], ptr [[TMP3]], i64 0, i64 [[IDXPROM7]]
@@ -4573,7 +4569,7 @@ int main() {
//
//
// CHECK4-LABEL: define {{[^@]+}}@__main_block_invoke
-// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR7:[0-9]+]] {
+// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca ptr, align 8
// CHECK4-NEXT: [[BLOCK_ADDR:%.*]] = alloca ptr, align 8
@@ -4603,14 +4599,14 @@ int main() {
// CHECK4-NEXT: [[DOTOMP_REDUCTION_RED_LIST:%.*]] = alloca [2 x ptr], align 8
// CHECK4-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK4-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
-// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr @g1, align 8, !nonnull [[META5:![0-9]+]], !align [[META6:![0-9]+]]
+// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr @g1, align 8, !nonnull [[META4:![0-9]+]], !align [[META5:![0-9]+]]
// CHECK4-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
// CHECK4-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK4-NEXT: store i32 1, ptr [[DOTOMP_UB]], align 4
// CHECK4-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK4-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK4-NEXT: store double 0.000000e+00, ptr [[G]], align 8
-// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr @g1, align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr @g1, align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store double 0.000000e+00, ptr [[G1]], align 8
// CHECK4-NEXT: store ptr [[G1]], ptr [[_TMP2]], align 8
// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
@@ -4641,7 +4637,7 @@ int main() {
// CHECK4-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK4-NEXT: store i32 [[ADD]], ptr [[I]], align 4
// CHECK4-NEXT: store double 1.000000e+00, ptr [[G]], align 8
-// CHECK4-NEXT: [[TMP10:%.*]] = load ptr, ptr [[_TMP2]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP10:%.*]] = load ptr, ptr [[_TMP2]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store volatile double 1.000000e+00, ptr [[TMP10]], align 8
// CHECK4-NEXT: [[BLOCK_ISA:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, double, ptr }>, ptr [[BLOCK]], i32 0, i32 0
// CHECK4-NEXT: store ptr @_NSConcreteStackBlock, ptr [[BLOCK_ISA]], align 8
@@ -4657,7 +4653,7 @@ int main() {
// CHECK4-NEXT: [[TMP11:%.*]] = load volatile double, ptr [[G]], align 8
// CHECK4-NEXT: store volatile double [[TMP11]], ptr [[BLOCK_CAPTURED]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURED4:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, double, ptr }>, ptr [[BLOCK]], i32 0, i32 6
-// CHECK4-NEXT: [[TMP12:%.*]] = load ptr, ptr [[_TMP2]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP12:%.*]] = load ptr, ptr [[_TMP2]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP12]], ptr [[BLOCK_CAPTURED4]], align 8
// CHECK4-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT___BLOCK_LITERAL_GENERIC:%.*]], ptr [[BLOCK]], i32 0, i32 3
// CHECK4-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
@@ -4707,7 +4703,7 @@ int main() {
//
//
// CHECK4-LABEL: define {{[^@]+}}@g1_block_invoke
-// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR7]] {
+// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca ptr, align 8
// CHECK4-NEXT: [[BLOCK_ADDR:%.*]] = alloca ptr, align 8
@@ -4716,7 +4712,7 @@ int main() {
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, double, ptr }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 5
// CHECK4-NEXT: store double 2.000000e+00, ptr [[BLOCK_CAPTURE_ADDR]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR1:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, double, ptr }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 6
-// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store double 2.000000e+00, ptr [[TMP0]], align 8
// CHECK4-NEXT: ret void
//
diff --git a/clang/test/OpenMP/for_reduction_codegen_UDR.cpp b/clang/test/OpenMP/for_reduction_codegen_UDR.cpp
index 0ac82685803db..08c90f0094f50 100644
--- a/clang/test/OpenMP/for_reduction_codegen_UDR.cpp
+++ b/clang/test/OpenMP/for_reduction_codegen_UDR.cpp
@@ -508,29 +508,29 @@ int main() {
// 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 [[S]], ptr [[S_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[S_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[S_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: call void @.omp_initializer.(ptr noundef [[S1]], ptr noundef [[TMP0]])
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP5:![0-9]+]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP4:![0-9]+]]
// CHECK1-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP1]], 10
// CHECK1-NEXT: br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP4]]
// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP2]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
-// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP5]]
+// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP4]]
// 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: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP4]]
// CHECK1-NEXT: [[ADD2:%.*]] = add nsw i32 [[TMP3]], 1
-// CHECK1-NEXT: store i32 [[ADD2]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP5]]
-// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP6:![0-9]+]]
+// CHECK1-NEXT: store i32 [[ADD2]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP4]]
+// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP5:![0-9]+]]
// CHECK1: omp.inner.for.end:
// CHECK1-NEXT: store i32 10, ptr [[I]], align 4
// CHECK1-NEXT: call void @.omp_combiner.(ptr noundef [[TMP0]], ptr noundef [[S1]])
@@ -654,7 +654,7 @@ int main() {
// CHECK1-NEXT: br i1 [[ARRAYCTOR_DONE8]], label [[ARRAYCTOR_CONT9:%.*]], label [[ARRAYCTOR_LOOP5]]
// CHECK1: arrayctor.cont9:
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[VAR3]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 6, ptr @main.omp_outlined, ptr [[T_VAR]], ptr [[TMP0]], ptr [[VAR1]], ptr [[T_VAR1]], ptr [[VEC]], ptr [[S_ARR]])
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 1
// CHECK1-NEXT: [[TMP1:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
@@ -668,9 +668,9 @@ int main() {
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @main.omp_outlined.11, i64 10, i64 [[TMP2]], ptr [[VLA]], ptr [[ARRS]])
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 1, ptr @main.omp_outlined.12, ptr [[VAR2]])
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 1, ptr @main.omp_outlined.13, ptr [[VVAR2]])
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 1, ptr @main.omp_outlined.14, ptr [[TMP5]])
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 1, ptr @main.omp_outlined.15, ptr [[TMP6]])
// CHECK1-NEXT: [[CALL10:%.*]] = call noundef i32 @_Z5tmainIiLi42EET_v()
// CHECK1-NEXT: store i32 [[CALL10]], ptr [[RETVAL]], align 4
@@ -769,21 +769,21 @@ int main() {
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VEC]], ptr [[VEC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_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: call void @.omp_initializer..2(ptr noundef [[T_VAR3]], ptr noundef [[TMP0]])
-// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[ADD_PTR:%.*]] = getelementptr inbounds i8, ptr [[VAR4]], i64 4
// CHECK1-NEXT: [[ADD_PTR5:%.*]] = getelementptr inbounds i8, ptr [[TMP7]], i64 4
// CHECK1-NEXT: call void @.omp_initializer..4(ptr noundef [[ADD_PTR]], ptr noundef [[ADD_PTR5]])
@@ -826,7 +826,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP18]] to i64
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP4]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4
-// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM10:%.*]] = sext i32 [[TMP20]] to i64
// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds [4 x [[STRUCT_S_0]]], ptr [[TMP5]], i64 0, i64 [[IDXPROM10]]
@@ -1065,9 +1065,9 @@ int main() {
// CHECK1-NEXT: store ptr [[ARRS]], ptr [[ARRS_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR2]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -1085,15 +1085,14 @@ int main() {
// CHECK1-NEXT: [[TMP9:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
// CHECK1-NEXT: [[TMP11:%.*]] = sub i64 [[TMP9]], [[TMP10]]
-// CHECK1-NEXT: [[TMP12:%.*]] = sdiv exact i64 [[TMP11]], 4
-// CHECK1-NEXT: [[TMP13:%.*]] = add nuw i64 [[TMP12]], 1
-// CHECK1-NEXT: [[TMP14:%.*]] = mul nuw i64 [[TMP13]], ptrtoint (ptr getelementptr (i32, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP15:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP15]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP12:%.*]] = add nuw i64 [[TMP11]], 4
+// CHECK1-NEXT: [[TMP13:%.*]] = udiv exact i64 [[TMP12]], 4
+// CHECK1-NEXT: [[TMP14:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP14]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA7:%.*]] = alloca i32, i64 [[TMP13]], align 16
// CHECK1-NEXT: store i64 [[TMP13]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP13]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA7]], [[TMP16]]
+// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA7]], [[TMP15]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[ARRAYIDX3]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
@@ -1101,34 +1100,33 @@ int main() {
// CHECK1-NEXT: call void @.omp_initializer..10(ptr noundef [[OMP_ARRAYCPY_DESTELEMENTPAST]], ptr noundef [[OMP_ARRAYCPY_SRCELEMENTPAST]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT8]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT8]], [[TMP16]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT8]], [[TMP15]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
-// CHECK1-NEXT: [[TMP20:%.*]] = sdiv exact i64 [[TMP19]], 4
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP20]]
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = sdiv exact i64 [[TMP18]], 4
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i32, ptr [[VLA7]], i64 [[TMP19]]
// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw [10 x [4 x [[STRUCT_S_0:%.*]]]], ptr [[TMP4]], i64 0, i64 1
// CHECK1-NEXT: [[ARRAYDECAY:%.*]] = getelementptr inbounds [4 x [[STRUCT_S_0]]], ptr [[ARRAYIDX9]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw [[STRUCT_S_0]], ptr [[ARRAYDECAY]], i64 1
// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP3]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[ARRAYIDX11]], align 4
-// CHECK1-NEXT: [[TMP23:%.*]] = sext i32 [[TMP22]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN12:%.*]] = add nsw i64 0, [[TMP23]]
+// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[ARRAYIDX11]], align 4
+// CHECK1-NEXT: [[TMP22:%.*]] = sext i32 [[TMP21]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN12:%.*]] = add nsw i64 0, [[TMP22]]
// CHECK1-NEXT: [[ARRAYIDX13:%.*]] = getelementptr inbounds nuw [10 x [4 x [[STRUCT_S_0]]]], ptr [[TMP4]], i64 0, i64 [[LB_ADD_LEN12]]
// CHECK1-NEXT: [[ARRAYDECAY14:%.*]] = getelementptr inbounds [4 x [[STRUCT_S_0]]], ptr [[ARRAYIDX13]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYIDX15:%.*]] = getelementptr inbounds nuw [[STRUCT_S_0]], ptr [[ARRAYDECAY14]], i64 2
-// CHECK1-NEXT: [[TMP24:%.*]] = ptrtoaddr ptr [[ARRAYIDX15]] to i64
-// CHECK1-NEXT: [[TMP25:%.*]] = ptrtoaddr ptr [[ARRAYIDX10]] to i64
-// CHECK1-NEXT: [[TMP26:%.*]] = sub i64 [[TMP24]], [[TMP25]]
-// CHECK1-NEXT: [[TMP27:%.*]] = sdiv exact i64 [[TMP26]], 12
-// CHECK1-NEXT: [[TMP28:%.*]] = add nuw i64 [[TMP27]], 1
-// CHECK1-NEXT: [[TMP29:%.*]] = mul nuw i64 [[TMP28]], ptrtoint (ptr getelementptr ([[STRUCT_S_0]], ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[VLA16:%.*]] = alloca [[STRUCT_S_0]], i64 [[TMP28]], align 16
-// CHECK1-NEXT: store i64 [[TMP28]], ptr [[__VLA_EXPR1]], align 8
-// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA16]], i64 [[TMP28]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY17:%.*]] = icmp eq ptr [[VLA16]], [[TMP30]]
+// CHECK1-NEXT: [[TMP23:%.*]] = ptrtoaddr ptr [[ARRAYIDX15]] to i64
+// CHECK1-NEXT: [[TMP24:%.*]] = ptrtoaddr ptr [[ARRAYIDX10]] to i64
+// CHECK1-NEXT: [[TMP25:%.*]] = sub i64 [[TMP23]], [[TMP24]]
+// CHECK1-NEXT: [[TMP26:%.*]] = add nuw i64 [[TMP25]], 12
+// CHECK1-NEXT: [[TMP27:%.*]] = udiv exact i64 [[TMP26]], 12
+// CHECK1-NEXT: [[VLA16:%.*]] = alloca [[STRUCT_S_0]], i64 [[TMP27]], align 16
+// CHECK1-NEXT: store i64 [[TMP27]], ptr [[__VLA_EXPR1]], align 8
+// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA16]], i64 [[TMP27]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY17:%.*]] = icmp eq ptr [[VLA16]], [[TMP28]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY17]], label [[OMP_ARRAYINIT_DONE25:%.*]], label [[OMP_ARRAYINIT_BODY18:%.*]]
// CHECK1: omp.arrayinit.body18:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST19:%.*]] = phi ptr [ [[ARRAYIDX10]], [[OMP_ARRAYINIT_DONE]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT22:%.*]], [[OMP_ARRAYINIT_BODY18]] ]
@@ -1138,85 +1136,85 @@ int main() {
// CHECK1-NEXT: call void @.omp_initializer..4(ptr noundef [[ADD_PTR]], ptr noundef [[ADD_PTR21]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT22]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST19]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT23]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST20]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE24:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT23]], [[TMP30]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE24:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT23]], [[TMP28]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE24]], label [[OMP_ARRAYINIT_DONE25]], label [[OMP_ARRAYINIT_BODY18]]
// CHECK1: omp.arrayinit.done25:
-// CHECK1-NEXT: [[TMP31:%.*]] = ptrtoaddr ptr [[TMP4]] to i64
-// CHECK1-NEXT: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX10]] to i64
-// CHECK1-NEXT: [[TMP33:%.*]] = sub i64 [[TMP31]], [[TMP32]]
-// CHECK1-NEXT: [[TMP34:%.*]] = sdiv exact i64 [[TMP33]], 12
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA16]], i64 [[TMP34]]
-// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[TMP36]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2]], i32 [[TMP37]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP38]], 9
+// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP4]] to i64
+// CHECK1-NEXT: [[TMP30:%.*]] = ptrtoaddr ptr [[ARRAYIDX10]] to i64
+// CHECK1-NEXT: [[TMP31:%.*]] = sub i64 [[TMP29]], [[TMP30]]
+// CHECK1-NEXT: [[TMP32:%.*]] = sdiv exact i64 [[TMP31]], 12
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA16]], i64 [[TMP32]]
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = load i32, ptr [[TMP34]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2]], i32 [[TMP35]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP36:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP36]], 9
// 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: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP39]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP37]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
-// CHECK1-NEXT: store i32 [[TMP40]], ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
+// CHECK1-NEXT: store i32 [[TMP38]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP26:%.*]] = icmp sle i32 [[TMP41]], [[TMP42]]
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP26:%.*]] = icmp sle i32 [[TMP39]], [[TMP40]]
// CHECK1-NEXT: br i1 [[CMP26]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP43]], 1
+// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP41]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], align 4
-// CHECK1-NEXT: [[TMP44:%.*]] = mul nsw i64 1, [[TMP1]]
-// CHECK1-NEXT: [[ARRAYIDX27:%.*]] = getelementptr inbounds i32, ptr [[TMP21]], i64 [[TMP44]]
-// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[I]], align 4
-// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP45]] to i64
+// CHECK1-NEXT: [[TMP42:%.*]] = mul nsw i64 1, [[TMP1]]
+// CHECK1-NEXT: [[ARRAYIDX27:%.*]] = getelementptr inbounds i32, ptr [[TMP20]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[I]], align 4
+// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP43]] to i64
// CHECK1-NEXT: [[ARRAYIDX28:%.*]] = getelementptr inbounds i32, ptr [[ARRAYIDX27]], i64 [[IDXPROM]]
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[ARRAYIDX28]], align 4
-// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP46]], 1
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[ARRAYIDX28]], align 4
+// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP44]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[ARRAYIDX28]], 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: [[TMP47:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[ADD29:%.*]] = add nsw i32 [[TMP47]], 1
+// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[ADD29:%.*]] = add nsw i32 [[TMP45]], 1
// CHECK1-NEXT: store i32 [[ADD29]], 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: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP49]])
-// CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA7]], ptr [[TMP50]], align 8
-// CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP52:%.*]] = inttoptr i64 [[TMP13]] to ptr
-// CHECK1-NEXT: store ptr [[TMP52]], ptr [[TMP51]], align 8
-// CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: store ptr [[VLA16]], ptr [[TMP53]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
-// CHECK1-NEXT: [[TMP55:%.*]] = inttoptr i64 [[TMP28]] to ptr
-// CHECK1-NEXT: store ptr [[TMP55]], ptr [[TMP54]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = load i32, ptr [[TMP56]], align 4
-// CHECK1-NEXT: [[TMP58:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP57]], i32 2, i64 32, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.8.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP58]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP47]])
+// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA7]], ptr [[TMP48]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP50:%.*]] = inttoptr i64 [[TMP13]] to ptr
+// CHECK1-NEXT: store ptr [[TMP50]], ptr [[TMP49]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: store ptr [[VLA16]], ptr [[TMP51]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = getelementptr inbounds [4 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 3
+// CHECK1-NEXT: [[TMP53:%.*]] = inttoptr i64 [[TMP27]] to ptr
+// CHECK1-NEXT: store ptr [[TMP53]], ptr [[TMP52]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = load i32, ptr [[TMP54]], align 4
+// CHECK1-NEXT: [[TMP56:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP55]], i32 2, i64 32, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.8.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP56]], 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: [[TMP59:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP59]]
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP57]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE34:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST30:%.*]] = phi ptr [ [[VLA7]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
@@ -1224,11 +1222,11 @@ int main() {
// CHECK1-NEXT: call void @.omp_combiner..9(ptr noundef [[OMP_ARRAYCPY_DESTELEMENTPAST31]], ptr noundef [[OMP_ARRAYCPY_SRCELEMENTPAST30]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT32]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST31]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST30]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE33:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT32]], [[TMP59]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE33:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT32]], [[TMP57]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE33]], label [[OMP_ARRAYCPY_DONE34]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done34:
-// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX10]], i64 [[TMP28]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY35:%.*]] = icmp eq ptr [[ARRAYIDX10]], [[TMP60]]
+// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX10]], i64 [[TMP27]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY35:%.*]] = icmp eq ptr [[ARRAYIDX10]], [[TMP58]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY35]], label [[OMP_ARRAYCPY_DONE44:%.*]], label [[OMP_ARRAYCPY_BODY36:%.*]]
// CHECK1: omp.arraycpy.body36:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST37:%.*]] = phi ptr [ [[VLA16]], [[OMP_ARRAYCPY_DONE34]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT42:%.*]], [[OMP_ARRAYCPY_BODY36]] ]
@@ -1238,60 +1236,60 @@ int main() {
// CHECK1-NEXT: call void @.omp_combiner..3(ptr noundef [[ADD_PTR39]], ptr noundef [[ADD_PTR40]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT41]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST38]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT42]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST37]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE43:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT41]], [[TMP60]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE43:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT41]], [[TMP58]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE43]], label [[OMP_ARRAYCPY_DONE44]], label [[OMP_ARRAYCPY_BODY36]]
// CHECK1: omp.arraycpy.done44:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP57]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP55]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP61:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY45:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP61]]
+// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr i32, ptr [[ARRAYIDX3]], i64 [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY45:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP59]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY45]], label [[OMP_ARRAYCPY_DONE52:%.*]], label [[OMP_ARRAYCPY_BODY46:%.*]]
// CHECK1: omp.arraycpy.body46:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST47:%.*]] = phi ptr [ [[VLA7]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT50:%.*]], [[OMP_ARRAYCPY_BODY46]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST48:%.*]] = phi ptr [ [[ARRAYIDX3]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT49:%.*]], [[OMP_ARRAYCPY_BODY46]] ]
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP63]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
+// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP61]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: call void @.omp_combiner..9(ptr noundef [[OMP_ARRAYCPY_DESTELEMENTPAST48]], ptr noundef [[OMP_ARRAYCPY_SRCELEMENTPAST47]])
-// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP63]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP61]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT49]] = getelementptr i32, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST48]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT50]] = getelementptr i32, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST47]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE51:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT49]], [[TMP61]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE51:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT49]], [[TMP59]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE51]], label [[OMP_ARRAYCPY_DONE52]], label [[OMP_ARRAYCPY_BODY46]]
// CHECK1: omp.arraycpy.done52:
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX10]], i64 [[TMP28]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY53:%.*]] = icmp eq ptr [[ARRAYIDX10]], [[TMP64]]
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX10]], i64 [[TMP27]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY53:%.*]] = icmp eq ptr [[ARRAYIDX10]], [[TMP62]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY53]], label [[OMP_ARRAYCPY_DONE62:%.*]], label [[OMP_ARRAYCPY_BODY54:%.*]]
// CHECK1: omp.arraycpy.body54:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST55:%.*]] = phi ptr [ [[VLA16]], [[OMP_ARRAYCPY_DONE52]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT60:%.*]], [[OMP_ARRAYCPY_BODY54]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST56:%.*]] = phi ptr [ [[ARRAYIDX10]], [[OMP_ARRAYCPY_DONE52]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT59:%.*]], [[OMP_ARRAYCPY_BODY54]] ]
-// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP66:%.*]] = load i32, ptr [[TMP65]], align 4
-// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP66]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: [[TMP63:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP64:%.*]] = load i32, ptr [[TMP63]], align 4
+// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP64]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[ADD_PTR57:%.*]] = getelementptr inbounds i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST56]], i64 4
// CHECK1-NEXT: [[ADD_PTR58:%.*]] = getelementptr inbounds i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST55]], i64 4
// CHECK1-NEXT: call void @.omp_combiner..3(ptr noundef [[ADD_PTR57]], ptr noundef [[ADD_PTR58]])
-// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP66]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP64]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT59]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST56]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT60]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST55]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE61:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT59]], [[TMP64]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE61:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT59]], [[TMP62]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE61]], label [[OMP_ARRAYCPY_DONE62]], label [[OMP_ARRAYCPY_BODY54]]
// CHECK1: omp.arraycpy.done62:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[VLA16]], i64 [[TMP28]]
-// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA16]], [[TMP67]]
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[VLA16]], i64 [[TMP27]]
+// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA16]], [[TMP65]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE63:%.*]], label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1: arraydestroy.body:
-// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP67]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
+// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP65]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT: call void @_ZN1SIfED1Ev(ptr noundef nonnull align 4 dead_on_return(12) dereferenceable(12) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]]
// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[VLA16]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE63]], label [[ARRAYDESTROY_BODY]]
// CHECK1: arraydestroy.done63:
-// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP68]])
+// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP66]])
// CHECK1-NEXT: ret void
//
//
@@ -1403,15 +1401,15 @@ int main() {
// CHECK1-NEXT: store ptr [[ARRS]], ptr [[ARRS_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR2]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[ARRS_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP4:%.*]] = mul nuw i64 [[TMP0]], [[TMP1]]
// CHECK1-NEXT: [[TMP5:%.*]] = mul nuw i64 [[TMP4]], 4
-// CHECK1-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], ptrtoint (ptr getelementptr (i32, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], 4
// CHECK1-NEXT: [[TMP7:%.*]] = call ptr @llvm.stacksave.p0()
// CHECK1-NEXT: store ptr [[TMP7]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA3:%.*]] = alloca i32, i64 [[TMP6]], align 16
@@ -1663,7 +1661,7 @@ int main() {
// 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 [[VAR2]], ptr [[VAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META3]], !align [[META11:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR2_ADDR]], align 8, !nonnull [[META2]], !align [[META10:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -1679,15 +1677,14 @@ int main() {
// CHECK1-NEXT: [[TMP5:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
// CHECK1-NEXT: [[TMP6:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
// CHECK1-NEXT: [[TMP7:%.*]] = sub i64 [[TMP5]], [[TMP6]]
-// CHECK1-NEXT: [[TMP8:%.*]] = sdiv exact i64 [[TMP7]], 12
-// CHECK1-NEXT: [[TMP9:%.*]] = add nuw i64 [[TMP8]], 1
-// CHECK1-NEXT: [[TMP10:%.*]] = mul nuw i64 [[TMP9]], ptrtoint (ptr getelementptr ([[STRUCT_S_0]], ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP8:%.*]] = add nuw i64 [[TMP7]], 12
+// CHECK1-NEXT: [[TMP9:%.*]] = udiv exact i64 [[TMP8]], 12
+// CHECK1-NEXT: [[TMP10:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP10]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca [[STRUCT_S_0]], i64 [[TMP9]], align 16
// CHECK1-NEXT: store i64 [[TMP9]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA]], i64 [[TMP9]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
+// CHECK1-NEXT: [[TMP11:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA]], i64 [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP11]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[ARRAYIDX1]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
@@ -1697,76 +1694,76 @@ int main() {
// CHECK1-NEXT: call void @.omp_initializer..4(ptr noundef [[ADD_PTR]], ptr noundef [[ADD_PTR4]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT5]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP11]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[TMP0]], align 8
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
-// CHECK1-NEXT: [[TMP18:%.*]] = sdiv exact i64 [[TMP17]], 12
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[TMP0]], align 8
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = ptrtoaddr ptr [[TMP13]] to i64
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = sub i64 [[TMP14]], [[TMP15]]
+// CHECK1-NEXT: [[TMP17:%.*]] = sdiv exact i64 [[TMP16]], 12
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP7]], ptr [[_TMP6]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP7]], align 8
-// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[TMP20]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2]], i32 [[TMP21]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP22]], 9
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP7]], align 8
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[TMP19]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2]], i32 [[TMP20]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP21:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP21]], 9
// 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: [[TMP23:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[TMP22:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP23]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i32 [ 9, [[COND_TRUE]] ], [ [[TMP22]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[TMP24:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
-// CHECK1-NEXT: store i32 [[TMP24]], ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
+// CHECK1-NEXT: store i32 [[TMP23]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP25:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[TMP26:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
-// CHECK1-NEXT: [[CMP8:%.*]] = icmp sle i32 [[TMP25]], [[TMP26]]
+// CHECK1-NEXT: [[TMP24:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[TMP25:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
+// CHECK1-NEXT: [[CMP8:%.*]] = icmp sle i32 [[TMP24]], [[TMP25]]
// CHECK1-NEXT: br i1 [[CMP8]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP27:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP27]], 1
+// CHECK1-NEXT: [[TMP26:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP26]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[I]], 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: [[TMP28:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
-// CHECK1-NEXT: [[ADD9:%.*]] = add nsw i32 [[TMP28]], 1
+// CHECK1-NEXT: [[TMP27:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
+// CHECK1-NEXT: [[ADD9:%.*]] = add nsw i32 [[TMP27]], 1
// CHECK1-NEXT: store i32 [[ADD9]], 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: [[TMP29:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP29]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP30]])
-// CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP31]], align 8
-// CHECK1-NEXT: [[TMP32:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP33:%.*]] = inttoptr i64 [[TMP9]] to ptr
-// CHECK1-NEXT: store ptr [[TMP33]], ptr [[TMP32]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = load i32, ptr [[TMP34]], align 4
-// CHECK1-NEXT: [[TMP36:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB3]], i32 [[TMP35]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.12.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP36]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP28]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP29]])
+// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP30]], align 8
+// CHECK1-NEXT: [[TMP31:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP32:%.*]] = inttoptr i64 [[TMP9]] to ptr
+// CHECK1-NEXT: store ptr [[TMP32]], ptr [[TMP31]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load i32, ptr [[TMP33]], align 4
+// CHECK1-NEXT: [[TMP35:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB3]], i32 [[TMP34]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.12.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP35]], 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: [[TMP37:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP37]]
+// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP36]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE16:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST10:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
@@ -1776,48 +1773,48 @@ int main() {
// CHECK1-NEXT: call void @.omp_combiner..3(ptr noundef [[ADD_PTR12]], ptr noundef [[ADD_PTR13]])
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT14]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE15:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT14]], [[TMP37]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE15:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT14]], [[TMP36]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE15]], label [[OMP_ARRAYCPY_DONE16]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done16:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP35]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP34]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY17:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP38]]
+// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAYIDX1]], i64 [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY17:%.*]] = icmp eq ptr [[ARRAYIDX1]], [[TMP37]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY17]], label [[OMP_ARRAYCPY_DONE26:%.*]], label [[OMP_ARRAYCPY_BODY18:%.*]]
// CHECK1: omp.arraycpy.body18:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST19:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT24:%.*]], [[OMP_ARRAYCPY_BODY18]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST20:%.*]] = phi ptr [ [[ARRAYIDX1]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT23:%.*]], [[OMP_ARRAYCPY_BODY18]] ]
-// CHECK1-NEXT: [[TMP39:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP39]], align 4
-// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP40]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[TMP38]], align 4
+// CHECK1-NEXT: call void @__kmpc_critical(ptr @[[GLOB1]], i32 [[TMP39]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[ADD_PTR21:%.*]] = getelementptr inbounds i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST20]], i64 4
// CHECK1-NEXT: [[ADD_PTR22:%.*]] = getelementptr inbounds i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST19]], i64 4
// CHECK1-NEXT: call void @.omp_combiner..3(ptr noundef [[ADD_PTR21]], ptr noundef [[ADD_PTR22]])
-// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP40]], ptr @.gomp_critical_user_.atomic_reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_critical(ptr @[[GLOB1]], i32 [[TMP39]], ptr @.gomp_critical_user_.atomic_reduction.var)
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT23]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST20]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT24]] = getelementptr [[STRUCT_S_0]], ptr [[OMP_ARRAYCPY_SRCELEMENTPAST19]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE25:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT23]], [[TMP38]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE25:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT23]], [[TMP37]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE25]], label [[OMP_ARRAYCPY_DONE26]], label [[OMP_ARRAYCPY_BODY18]]
// CHECK1: omp.arraycpy.done26:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP35]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP34]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[VLA]], i64 [[TMP9]]
-// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP41]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[VLA]], i64 [[TMP9]]
+// CHECK1-NEXT: [[ARRAYDESTROY_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP40]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_ISEMPTY]], label [[ARRAYDESTROY_DONE27:%.*]], label [[ARRAYDESTROY_BODY:%.*]]
// CHECK1: arraydestroy.body:
-// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP41]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
+// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENTPAST:%.*]] = phi ptr [ [[TMP40]], [[DOTOMP_REDUCTION_DEFAULT]] ], [ [[ARRAYDESTROY_ELEMENT:%.*]], [[ARRAYDESTROY_BODY]] ]
// CHECK1-NEXT: [[ARRAYDESTROY_ELEMENT]] = getelementptr inbounds [[STRUCT_S_0]], ptr [[ARRAYDESTROY_ELEMENTPAST]], i64 -1
// CHECK1-NEXT: call void @_ZN1SIfED1Ev(ptr noundef nonnull align 4 dead_on_return(12) dereferenceable(12) [[ARRAYDESTROY_ELEMENT]]) #[[ATTR3]]
// CHECK1-NEXT: [[ARRAYDESTROY_DONE:%.*]] = icmp eq ptr [[ARRAYDESTROY_ELEMENT]], [[VLA]]
// CHECK1-NEXT: br i1 [[ARRAYDESTROY_DONE]], label [[ARRAYDESTROY_DONE27]], label [[ARRAYDESTROY_BODY]]
// CHECK1: arraydestroy.done27:
-// CHECK1-NEXT: [[TMP42:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP42]])
-// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
-// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB4]], i32 [[TMP44]])
+// CHECK1-NEXT: [[TMP41:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP41]])
+// CHECK1-NEXT: [[TMP42:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[TMP42]], align 4
+// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB4]], i32 [[TMP43]])
// CHECK1-NEXT: ret void
//
//
@@ -1872,7 +1869,7 @@ int main() {
// 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 [[VVAR2]], ptr [[VVAR2_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VVAR2_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VVAR2_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_STRIDE]], align 4
@@ -2057,17 +2054,17 @@ int main() {
// 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 [[VAR3]], ptr [[VAR3_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S_0]]], ptr [[TMP2]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds nuw [4 x [[STRUCT_S_0]]], ptr [[TMP3]], i64 0, i64 2
// CHECK1-NEXT: [[ARRAY_BEGIN:%.*]] = getelementptr inbounds [2 x [[STRUCT_S_0]]], ptr [[VAR34]], i32 0, i32 0
// CHECK1-NEXT: [[TMP4:%.*]] = getelementptr [[STRUCT_S_0]], ptr [[ARRAY_BEGIN]], i64 2
@@ -2084,7 +2081,7 @@ int main() {
// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT6]], [[TMP4]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP6:%.*]] = ptrtoaddr ptr [[TMP5]] to i64
// CHECK1-NEXT: [[TMP7:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK1-NEXT: [[TMP8:%.*]] = sub i64 [[TMP6]], [[TMP7]]
@@ -2248,15 +2245,15 @@ int main() {
// 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 [[VAR3]], ptr [[VAR3_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR3_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 9, ptr [[DOTOMP_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: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[TMP3]], align 4
// CHECK1-NEXT: [[DOTVAR3__VOID_ADDR:%.*]] = call ptr @__kmpc_alloc(i32 [[TMP4]], i64 48, ptr inttoptr (i64 6 to ptr))
@@ -2435,11 +2432,11 @@ int main() {
// CHECK1-NEXT: [[ARRAYCTOR_DONE:%.*]] = icmp eq ptr [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK1-NEXT: br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK1: arrayctor.cont:
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 6, ptr @_Z5tmainIiLi42EET_v.omp_outlined, ptr [[T_VAR]], ptr [[TMP0]], ptr [[VAR1]], ptr [[T_VAR1]], ptr [[VEC]], ptr [[S_ARR]])
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @_Z5tmainIiLi42EET_v.omp_outlined.22, ptr [[T_VAR]], ptr [[VEC]], ptr [[S_ARR]], ptr [[TMP1]])
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 5, ptr @_Z5tmainIiLi42EET_v.omp_outlined.25, ptr [[ARR]], ptr [[VEC]], ptr [[T_VAR]], ptr [[S_ARR]], ptr [[TMP2]])
// CHECK1-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK1-NEXT: [[ARRAY_BEGIN1:%.*]] = getelementptr inbounds [42 x [[STRUCT_S]]], ptr [[ARR]], i32 0, i32 0
@@ -2557,21 +2554,21 @@ int main() {
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VEC]], ptr [[VEC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP6]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_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: call void @.omp_initializer..17(ptr noundef [[T_VAR3]], ptr noundef [[TMP0]])
-// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[ADD_PTR:%.*]] = getelementptr inbounds i8, ptr [[VAR4]], i64 4
// CHECK1-NEXT: [[ADD_PTR5:%.*]] = getelementptr inbounds i8, ptr [[TMP7]], i64 4
// CHECK1-NEXT: call void @.omp_initializer..4(ptr noundef [[ADD_PTR]], ptr noundef [[ADD_PTR5]])
@@ -2612,7 +2609,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP17]] to i64
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP4]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[TMP16]], ptr [[ARRAYIDX]], align 4
-// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP19:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM10:%.*]] = sext i32 [[TMP19]] to i64
// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[TMP5]], i64 0, i64 [[IDXPROM10]]
@@ -2823,12 +2820,12 @@ int main() {
// CHECK1-NEXT: store ptr [[VEC]], ptr [[VEC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP3]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP4]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_UB]], align 4
@@ -2868,7 +2865,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP15]] to i64
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP1]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[TMP14]], ptr [[ARRAYIDX]], align 4
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP17:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM5:%.*]] = sext i32 [[TMP17]] to i64
// CHECK1-NEXT: [[ARRAYIDX6:%.*]] = getelementptr inbounds [2 x [[STRUCT_S:%.*]]], ptr [[TMP2]], i64 0, i64 [[IDXPROM5]]
@@ -2967,13 +2964,13 @@ int main() {
// CHECK1-NEXT: store ptr [[T_VAR]], ptr [[T_VAR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[S_ARR]], ptr [[S_ARR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP4]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: store ptr [[TMP5]], ptr [[_TMP1]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_UB]], align 4
@@ -3035,7 +3032,7 @@ int main() {
// CHECK1-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP21]] to i64
// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds [2 x i32], ptr [[TMP1]], i64 0, i64 [[IDXPROM]]
// CHECK1-NEXT: store i32 [[TMP20]], ptr [[ARRAYIDX8]], align 4
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[I]], align 4
// CHECK1-NEXT: [[IDXPROM9:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[TMP3]], i64 0, i64 [[IDXPROM9]]
@@ -3206,24 +3203,24 @@ int main() {
// CHECK3-NEXT: call void @.omp_initializer.(ptr noundef [[S1]], ptr noundef [[S]])
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3: omp.inner.for.cond:
-// CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2:![0-9]+]]
+// CHECK3-NEXT: [[TMP0:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1:![0-9]+]]
// CHECK3-NEXT: [[CMP:%.*]] = icmp slt i32 [[TMP0]], 10
// CHECK3-NEXT: br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK3: omp.inner.for.cond.cleanup:
// CHECK3-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK3: omp.inner.for.body:
-// CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP1:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP1]], 1
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
-// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP1]]
// 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: [[TMP2:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP2:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[ADD2:%.*]] = add nsw i32 [[TMP2]], 1
-// CHECK3-NEXT: store i32 [[ADD2]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
-// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
+// CHECK3-NEXT: store i32 [[ADD2]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
+// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP2:![0-9]+]]
// CHECK3: omp.inner.for.end:
// CHECK3-NEXT: store i32 10, ptr [[I]], align 4
// CHECK3-NEXT: call void @.omp_combiner.(ptr noundef [[S]], ptr noundef [[S1]])
@@ -3348,9 +3345,9 @@ int main() {
// CHECK3-NEXT: br i1 [[ARRAYCTOR_DONE8]], label [[ARRAYCTOR_CONT9:%.*]], label [[ARRAYCTOR_LOOP5]]
// CHECK3: arrayctor.cont9:
// CHECK3-NEXT: store ptr [[S_ARR]], ptr [[VAR3]], align 8
-// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META6:![0-9]+]], !align [[META7:![0-9]+]]
+// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5:![0-9]+]], !align [[META6:![0-9]+]]
// CHECK3-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[_TMP10]], align 8
// CHECK3-NEXT: store i32 0, ptr [[I]], align 4
// CHECK3-NEXT: br label [[FOR_COND:%.*]]
@@ -3365,7 +3362,7 @@ int main() {
// CHECK3-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP4]] to i64
// CHECK3-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 [[IDXPROM]]
// CHECK3-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4
-// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP10]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP10]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[I]], align 4
// CHECK3-NEXT: [[IDXPROM11:%.*]] = sext i32 [[TMP6]] to i64
// CHECK3-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds [4 x [[STRUCT_S_0]]], ptr [[S_ARR]], i64 0, i64 [[IDXPROM11]]
@@ -3375,7 +3372,7 @@ int main() {
// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[I]], align 4
// CHECK3-NEXT: [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK3-NEXT: store i32 [[INC]], ptr [[I]], align 4
-// CHECK3-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP8:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP7:![0-9]+]]
// CHECK3: for.end:
// CHECK3-NEXT: [[ARRAYIDX14:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 1
// CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[ARRAYIDX14]], align 4
@@ -3405,7 +3402,7 @@ int main() {
// CHECK3-NEXT: [[TMP16:%.*]] = load i32, ptr [[I15]], align 4
// CHECK3-NEXT: [[INC24:%.*]] = add nsw i32 [[TMP16]], 1
// CHECK3-NEXT: store i32 [[INC24]], ptr [[I15]], align 4
-// CHECK3-NEXT: br label [[FOR_COND16]], !llvm.loop [[LOOP10:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND16]], !llvm.loop [[LOOP9:![0-9]+]]
// CHECK3: for.end25:
// CHECK3-NEXT: store i32 0, ptr [[I26]], align 4
// CHECK3-NEXT: br label [[FOR_COND27:%.*]]
@@ -3427,7 +3424,7 @@ int main() {
// CHECK3-NEXT: [[TMP21:%.*]] = load i32, ptr [[I26]], align 4
// CHECK3-NEXT: [[INC35:%.*]] = add nsw i32 [[TMP21]], 1
// CHECK3-NEXT: store i32 [[INC35]], ptr [[I26]], align 4
-// CHECK3-NEXT: br label [[FOR_COND27]], !llvm.loop [[LOOP11:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND27]], !llvm.loop [[LOOP10:![0-9]+]]
// CHECK3: for.end36:
// CHECK3-NEXT: store i32 0, ptr [[I37]], align 4
// CHECK3-NEXT: br label [[FOR_COND38:%.*]]
@@ -3441,7 +3438,7 @@ int main() {
// CHECK3-NEXT: [[TMP23:%.*]] = load i32, ptr [[I37]], align 4
// CHECK3-NEXT: [[INC42:%.*]] = add nsw i32 [[TMP23]], 1
// CHECK3-NEXT: store i32 [[INC42]], ptr [[I37]], align 4
-// CHECK3-NEXT: br label [[FOR_COND38]], !llvm.loop [[LOOP12:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND38]], !llvm.loop [[LOOP11:![0-9]+]]
// CHECK3: for.end43:
// CHECK3-NEXT: store i32 0, ptr [[I44]], align 4
// CHECK3-NEXT: br label [[FOR_COND45:%.*]]
@@ -3455,11 +3452,11 @@ int main() {
// CHECK3-NEXT: [[TMP25:%.*]] = load i32, ptr [[I44]], align 4
// CHECK3-NEXT: [[INC49:%.*]] = add nsw i32 [[TMP25]], 1
// CHECK3-NEXT: store i32 [[INC49]], ptr [[I44]], align 4
-// CHECK3-NEXT: br label [[FOR_COND45]], !llvm.loop [[LOOP13:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND45]], !llvm.loop [[LOOP12:![0-9]+]]
// CHECK3: for.end50:
-// CHECK3-NEXT: [[TMP26:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP26:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP26]], ptr [[_TMP51]], align 8
-// CHECK3-NEXT: [[TMP27:%.*]] = load ptr, ptr [[_TMP51]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP27:%.*]] = load ptr, ptr [[_TMP51]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP27]], ptr [[_TMP52]], align 8
// CHECK3-NEXT: store i32 0, ptr [[I53]], align 4
// CHECK3-NEXT: br label [[FOR_COND54:%.*]]
@@ -3473,11 +3470,11 @@ int main() {
// CHECK3-NEXT: [[TMP29:%.*]] = load i32, ptr [[I53]], align 4
// CHECK3-NEXT: [[INC58:%.*]] = add nsw i32 [[TMP29]], 1
// CHECK3-NEXT: store i32 [[INC58]], ptr [[I53]], align 4
-// CHECK3-NEXT: br label [[FOR_COND54]], !llvm.loop [[LOOP14:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND54]], !llvm.loop [[LOOP13:![0-9]+]]
// CHECK3: for.end59:
-// CHECK3-NEXT: [[TMP30:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP30:%.*]] = load ptr, ptr [[VAR3]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP30]], ptr [[_TMP60]], align 8
-// CHECK3-NEXT: [[TMP31:%.*]] = load ptr, ptr [[_TMP60]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP31:%.*]] = load ptr, ptr [[_TMP60]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP31]], ptr [[_TMP61]], align 8
// CHECK3-NEXT: store i32 0, ptr [[I62]], align 4
// CHECK3-NEXT: br label [[FOR_COND63:%.*]]
@@ -3491,7 +3488,7 @@ int main() {
// CHECK3-NEXT: [[TMP33:%.*]] = load i32, ptr [[I62]], align 4
// CHECK3-NEXT: [[INC67:%.*]] = add nsw i32 [[TMP33]], 1
// CHECK3-NEXT: store i32 [[INC67]], ptr [[I62]], align 4
-// CHECK3-NEXT: br label [[FOR_COND63]], !llvm.loop [[LOOP15:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND63]], !llvm.loop [[LOOP14:![0-9]+]]
// CHECK3: for.end68:
// CHECK3-NEXT: [[CALL69:%.*]] = call noundef i32 @_Z5tmainIiLi42EET_v()
// CHECK3-NEXT: store i32 [[CALL69]], ptr [[RETVAL]], align 4
@@ -3596,9 +3593,9 @@ int main() {
// CHECK3-NEXT: [[ARRAYCTOR_DONE:%.*]] = icmp eq ptr [[ARRAYCTOR_NEXT]], [[ARRAYCTOR_END]]
// CHECK3-NEXT: br i1 [[ARRAYCTOR_DONE]], label [[ARRAYCTOR_CONT:%.*]], label [[ARRAYCTOR_LOOP]]
// CHECK3: arrayctor.cont:
-// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP0]], ptr [[TMP]], align 8
-// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[_TMP1]], align 8
// CHECK3-NEXT: store i32 0, ptr [[I]], align 4
// CHECK3-NEXT: br label [[FOR_COND:%.*]]
@@ -3612,7 +3609,7 @@ int main() {
// CHECK3-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP4]] to i64
// CHECK3-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 [[IDXPROM]]
// CHECK3-NEXT: store i32 [[TMP3]], ptr [[ARRAYIDX]], align 4
-// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[I]], align 4
// CHECK3-NEXT: [[IDXPROM2:%.*]] = sext i32 [[TMP6]] to i64
// CHECK3-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[S_ARR]], i64 0, i64 [[IDXPROM2]]
@@ -3622,11 +3619,11 @@ int main() {
// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[I]], align 4
// CHECK3-NEXT: [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK3-NEXT: store i32 [[INC]], ptr [[I]], align 4
-// CHECK3-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP16:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP15:![0-9]+]]
// CHECK3: for.end:
-// CHECK3-NEXT: [[TMP8:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP8:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP8]], ptr [[_TMP4]], align 8
-// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP4]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP4]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP9]], ptr [[_TMP5]], align 8
// CHECK3-NEXT: store i32 0, ptr [[I6]], align 4
// CHECK3-NEXT: br label [[FOR_COND7:%.*]]
@@ -3640,7 +3637,7 @@ int main() {
// CHECK3-NEXT: [[IDXPROM10:%.*]] = sext i32 [[TMP12]] to i64
// CHECK3-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 [[IDXPROM10]]
// CHECK3-NEXT: store i32 [[TMP11]], ptr [[ARRAYIDX11]], align 4
-// CHECK3-NEXT: [[TMP13:%.*]] = load ptr, ptr [[_TMP5]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP13:%.*]] = load ptr, ptr [[_TMP5]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: [[TMP14:%.*]] = load i32, ptr [[I6]], align 4
// CHECK3-NEXT: [[IDXPROM12:%.*]] = sext i32 [[TMP14]] to i64
// CHECK3-NEXT: [[ARRAYIDX13:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[S_ARR]], i64 0, i64 [[IDXPROM12]]
@@ -3650,11 +3647,11 @@ int main() {
// CHECK3-NEXT: [[TMP15:%.*]] = load i32, ptr [[I6]], align 4
// CHECK3-NEXT: [[INC16:%.*]] = add nsw i32 [[TMP15]], 1
// CHECK3-NEXT: store i32 [[INC16]], ptr [[I6]], align 4
-// CHECK3-NEXT: br label [[FOR_COND7]], !llvm.loop [[LOOP17:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND7]], !llvm.loop [[LOOP16:![0-9]+]]
// CHECK3: for.end17:
-// CHECK3-NEXT: [[TMP16:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP16:%.*]] = load ptr, ptr [[VAR]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP16]], ptr [[_TMP18]], align 8
-// CHECK3-NEXT: [[TMP17:%.*]] = load ptr, ptr [[_TMP18]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP17:%.*]] = load ptr, ptr [[_TMP18]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: store ptr [[TMP17]], ptr [[_TMP19]], align 8
// CHECK3-NEXT: store i32 0, ptr [[I20]], align 4
// CHECK3-NEXT: br label [[FOR_COND21:%.*]]
@@ -3668,7 +3665,7 @@ int main() {
// CHECK3-NEXT: [[IDXPROM24:%.*]] = sext i32 [[TMP20]] to i64
// CHECK3-NEXT: [[ARRAYIDX25:%.*]] = getelementptr inbounds [2 x i32], ptr [[VEC]], i64 0, i64 [[IDXPROM24]]
// CHECK3-NEXT: store i32 [[TMP19]], ptr [[ARRAYIDX25]], align 4
-// CHECK3-NEXT: [[TMP21:%.*]] = load ptr, ptr [[_TMP19]], align 8, !nonnull [[META6]], !align [[META7]]
+// CHECK3-NEXT: [[TMP21:%.*]] = load ptr, ptr [[_TMP19]], align 8, !nonnull [[META5]], !align [[META6]]
// CHECK3-NEXT: [[TMP22:%.*]] = load i32, ptr [[I20]], align 4
// CHECK3-NEXT: [[IDXPROM26:%.*]] = sext i32 [[TMP22]] to i64
// CHECK3-NEXT: [[ARRAYIDX27:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[S_ARR]], i64 0, i64 [[IDXPROM26]]
@@ -3678,7 +3675,7 @@ int main() {
// CHECK3-NEXT: [[TMP23:%.*]] = load i32, ptr [[I20]], align 4
// CHECK3-NEXT: [[INC30:%.*]] = add nsw i32 [[TMP23]], 1
// CHECK3-NEXT: store i32 [[INC30]], ptr [[I20]], align 4
-// CHECK3-NEXT: br label [[FOR_COND21]], !llvm.loop [[LOOP18:![0-9]+]]
+// CHECK3-NEXT: br label [[FOR_COND21]], !llvm.loop [[LOOP17:![0-9]+]]
// CHECK3: for.end31:
// CHECK3-NEXT: store i32 0, ptr [[RETVAL]], align 4
// CHECK3-NEXT: [[ARRAY_BEGIN32:%.*]] = getelementptr inbounds [42 x [[STRUCT_S]]], ptr [[ARR]], i32 0, i32 0
diff --git a/clang/test/OpenMP/for_reduction_task_codegen.cpp b/clang/test/OpenMP/for_reduction_task_codegen.cpp
index 0d3e1d28ba372..9710f051ae5e1 100644
--- a/clang/test/OpenMP/for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/for_reduction_task_codegen.cpp
@@ -60,8 +60,8 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
// CHECK1-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8
// CHECK1-NEXT: store i64 9, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: store i64 1, ptr [[DOTOMP_STRIDE]], align 8
@@ -82,227 +82,225 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP9:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP10:%.*]] = sub i64 [[TMP8]], [[TMP9]]
// CHECK1-NEXT: [[TMP11:%.*]] = add nuw i64 [[TMP10]], 1
-// CHECK1-NEXT: [[TMP12:%.*]] = mul nuw i64 [[TMP11]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP13:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP13]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP11]], align 16
// CHECK1-NEXT: store i64 [[TMP11]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP11]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP14]]
+// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP14]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP1]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP16]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP19]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP1]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
// CHECK1-NEXT: store ptr [[_TMP6]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP20]], ptr [[_TMP6]], align 8
+// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP6]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP26]], align 8
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP27]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_7:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[TMP1]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
-// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP30]], i64 0
-// CHECK1-NEXT: [[TMP31:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP32:%.*]] = sext i32 [[TMP31]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP32]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP1]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP33]], i64 9
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
-// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP34]], i64 [[LB_ADD_LEN10]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP28]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP38:%.*]] = sub i64 [[TMP36]], [[TMP37]]
-// CHECK1-NEXT: [[TMP39:%.*]] = add nuw i64 [[TMP38]], 1
-// CHECK1-NEXT: [[TMP40:%.*]] = mul nuw i64 [[TMP39]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP40]], ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP45]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
-// CHECK1-NEXT: [[TMP48:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP47]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP48]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = load i32, ptr [[TMP49]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP50]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP51]], 9
+// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[TMP1]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
+// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
+// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP31]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP1]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
+// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN10]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP34]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
+// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP38]], ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP43]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[TMP44]], align 4
+// CHECK1-NEXT: [[TMP46:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP45]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP46]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = load i32, ptr [[TMP47]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP48]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP49:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP49]], 9
// 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: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP52]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP50]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP53]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP51]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP54]], [[TMP55]]
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP52]], [[TMP53]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP56]], 1
+// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP54]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP57]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP60]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load i32, ptr [[TMP61]], align 4
-// CHECK1-NEXT: [[TMP63:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP62]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP63]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP64]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[TMP65]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP66]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP63]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP67]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP69]], ptr [[TMP68]], align 8
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[TMP70]], align 4
-// CHECK1-NEXT: [[TMP72:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP71]], ptr [[TMP63]])
+// CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP55]], align 8
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP58:%.*]] = load ptr, ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP58]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load i32, ptr [[TMP59]], align 4
+// CHECK1-NEXT: [[TMP61:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP60]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP61]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP62]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP64:%.*]] = load ptr, ptr [[TMP63]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP64]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP61]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP65]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP67]], ptr [[TMP66]], align 8
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
+// CHECK1-NEXT: [[TMP70:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP69]], ptr [[TMP61]])
// 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: [[TMP73:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP73]], 1
+// CHECK1-NEXT: [[TMP71:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP71]], 1
// CHECK1-NEXT: store i64 [[ADD14]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP72:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP73:%.*]] = load i32, ptr [[TMP72]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP73]])
// CHECK1-NEXT: [[TMP74:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP75:%.*]] = load i32, ptr [[TMP74]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP75]])
-// CHECK1-NEXT: [[TMP76:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP77:%.*]] = load i32, ptr [[TMP76]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP77]], i32 1)
-// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP78]], align 8
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP81:%.*]] = inttoptr i64 [[TMP11]] to ptr
-// CHECK1-NEXT: store ptr [[TMP81]], ptr [[TMP80]], align 8
-// CHECK1-NEXT: [[TMP82:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP83:%.*]] = load i32, ptr [[TMP82]], align 4
-// CHECK1-NEXT: [[TMP84:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB3:[0-9]+]], i32 [[TMP83]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP84]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP75]], i32 1)
+// CHECK1-NEXT: [[TMP76:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP76]], align 8
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP79:%.*]] = inttoptr i64 [[TMP11]] to ptr
+// CHECK1-NEXT: store ptr [[TMP79]], ptr [[TMP78]], align 8
+// CHECK1-NEXT: [[TMP80:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP81:%.*]] = load i32, ptr [[TMP80]], align 4
+// CHECK1-NEXT: [[TMP82:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB3:[0-9]+]], i32 [[TMP81]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP82]], 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: [[TMP85:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP86:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP85]], [[TMP86]]
+// CHECK1-NEXT: [[TMP83:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP83]], [[TMP84]]
// CHECK1-NEXT: store i32 [[ADD15]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP87:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP87]]
+// CHECK1-NEXT: [[TMP85:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP85]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE22:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST16:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT20:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP88]] to i32
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP89]] to i32
+// CHECK1-NEXT: [[TMP86:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP86]] to i32
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP87]] to i32
// CHECK1-NEXT: [[ADD18:%.*]] = add nsw i32 [[CONV]], [[CONV17]]
// CHECK1-NEXT: [[CONV19:%.*]] = trunc i32 [[ADD18]] to i8
// CHECK1-NEXT: store i8 [[CONV19]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT20]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP87]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP85]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_DONE22]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done22:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP83]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP81]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP90:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP91:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP90]] monotonic, align 4
-// CHECK1-NEXT: [[TMP92:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP92]]
+// CHECK1-NEXT: [[TMP88:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP89:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP88]] monotonic, align 4
+// CHECK1-NEXT: [[TMP90:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP90]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY23]], label [[OMP_ARRAYCPY_DONE36:%.*]], label [[OMP_ARRAYCPY_BODY24:%.*]]
// CHECK1: omp.arraycpy.body24:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST25:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT34:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT33:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP93:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP93]] to i32
+// CHECK1-NEXT: [[TMP91:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP91]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP94:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP99:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP94]], ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP95]] to i32
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP96]] to i32
+// CHECK1-NEXT: [[TMP92:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP97:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP92]], ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[TMP93:%.*]] = load i8, ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP93]] to i32
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP94]] to i32
// CHECK1-NEXT: [[ADD31:%.*]] = add nsw i32 [[CONV29]], [[CONV30]]
// CHECK1-NEXT: [[CONV32:%.*]] = trunc i32 [[ADD31]] to i8
// CHECK1-NEXT: store i8 [[CONV32]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP97:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP98:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP94]], i8 [[TMP97]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP99]] = extractvalue { i8, i1 } [[TMP98]], 0
-// CHECK1-NEXT: [[TMP100:%.*]] = extractvalue { i8, i1 } [[TMP98]], 1
-// CHECK1-NEXT: br i1 [[TMP100]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP96:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP92]], i8 [[TMP95]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP97]] = extractvalue { i8, i1 } [[TMP96]], 0
+// CHECK1-NEXT: [[TMP98:%.*]] = extractvalue { i8, i1 } [[TMP96]], 1
+// CHECK1-NEXT: br i1 [[TMP98]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT34]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP92]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP90]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_DONE36]], label [[OMP_ARRAYCPY_BODY24]]
// CHECK1: omp.arraycpy.done36:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP83]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP81]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP101:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP101]])
-// CHECK1-NEXT: [[TMP102:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP103:%.*]] = load i32, ptr [[TMP102]], align 4
-// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB4:[0-9]+]], i32 [[TMP103]])
+// CHECK1-NEXT: [[TMP99:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP99]])
+// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP101:%.*]] = load i32, ptr [[TMP100]], align 4
+// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB4:[0-9]+]], i32 [[TMP101]])
// CHECK1-NEXT: ret void
//
//
@@ -426,31 +424,31 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META15:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]]
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]], !inline_history [[META15:![0-9]+]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META14]]
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14]]
// CHECK1-NEXT: [[TMP17:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP15]], ptr [[TMP14]])
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[TMP18]], align 8
// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[TMP22]], align 4
// CHECK1-NEXT: [[TMP24:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[LB_ADD_LEN_I:%.*]] = add nsw i64 -1, [[TMP24]]
@@ -463,19 +461,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
-// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
+// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META14]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/master_taskloop_in_reduction_codegen.cpp b/clang/test/OpenMP/master_taskloop_in_reduction_codegen.cpp
index 7b7dff4b92d9e..3c9a5ba278866 100644
--- a/clang/test/OpenMP/master_taskloop_in_reduction_codegen.cpp
+++ b/clang/test/OpenMP/master_taskloop_in_reduction_codegen.cpp
@@ -145,7 +145,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP35:%.*]] = mul nuw i64 [[TMP2]], 2
-// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 2
// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP37]], align 8
// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
@@ -359,7 +359,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: call void @_ZN1SC1ERKS_(ptr nonnull align 4 dereferenceable(4) [[AGG_RESULT]], ptr nonnull align 4 dereferenceable(4) [[TMP0]])
// CHECK1-NEXT: ret void
//
@@ -451,11 +451,11 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[D]], ptr [[D_ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_1]], ptr [[DOTTASK_RED__ADDR2]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META3]], !align [[META6:![0-9]+]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META3]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META2]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META2]], !align [[META5]]
// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK1-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_master(ptr @[[GLOB1]], i32 [[TMP6]])
@@ -560,66 +560,65 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 8
// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 9
// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[TMP17]], align 8
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META14:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META16:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META15:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP19]], i32 0, i32 1
// CHECK1-NEXT: [[TMP21:%.*]] = load i64, ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
-// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]], !inline_history [[META18:![0-9]+]]
+// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: [[TMP29:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP27]], ptr [[TMP26]])
// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP19]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META3]], !align [[META5]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META2]], !align [[META4]]
// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP21]], 2
-// CHECK1-NEXT: [[TMP33:%.*]] = udiv exact i64 [[TMP32]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
-// CHECK1-NEXT: store i64 [[TMP33]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP35]], ptr [[TMP31]])
-// CHECK1-NEXT: [[TMP37:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP37]] to i32
-// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
+// CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP34]], ptr [[TMP31]])
+// CHECK1-NEXT: [[TMP36:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP36]] to i32
+// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1: omp.inner.for.cond.i:
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP39]]
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP38]]
// CHECK1-NEXT: br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__EXIT:%.*]]
// CHECK1: omp.inner.for.body.i:
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: store i32 [[TMP40]], ptr [[I_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[TMP29]], align 4
-// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP41]] to i64
-// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP36]], i64 [[IDXPROM_I]]
-// CHECK1-NEXT: [[TMP42:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2
-// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP42]] to i32
-// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[TMP29]], align 4
-// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP43]], [[CONV3_I]]
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: store i32 [[TMP39]], ptr [[I_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP29]], align 4
+// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP40]] to i64
+// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP35]], i64 [[IDXPROM_I]]
+// CHECK1-NEXT: [[TMP41:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2
+// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP41]] to i32
+// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[TMP29]], align 4
+// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP42]], [[CONV3_I]]
// CHECK1-NEXT: store i32 [[ADD4_I]], ptr [[TMP29]], align 4
-// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP44]], 1
-// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP43]], 1
+// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I]]
// CHECK1: .omp_outlined..exit:
// CHECK1-NEXT: ret i32 0
diff --git a/clang/test/OpenMP/master_taskloop_reduction_codegen.cpp b/clang/test/OpenMP/master_taskloop_reduction_codegen.cpp
index 8c3e3cf973454..4ee4d4934c2de 100644
--- a/clang/test/OpenMP/master_taskloop_reduction_codegen.cpp
+++ b/clang/test/OpenMP/master_taskloop_reduction_codegen.cpp
@@ -94,10 +94,9 @@ sum = 0.0;
// CHECK-DAG: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK-DAG: [[TMP33:%.*]] = ptrtoaddr ptr [[ARRAYIDX5]] to i64
// CHECK-DAG: [[TMP34:%.*]] = sub i64 [[TMP32]], [[TMP33]]
-// CHECK-DAG: [[TMP35:%.*]] = sdiv exact i64 [[TMP34]], 4
-// CHECK-DAG: [[TMP36:%.*]] = add nuw i64 [[TMP35]], 1
-// CHECK-DAG: [[TMP37:%.*]] = mul nuw i64 [[TMP36]], ptrtoint (ptr getelementptr (%struct.S, ptr null, i32 1) to i64)
-// CHECK-DAG: store i64 [[TMP37]], ptr [[TMP38:%[^,]+]],
+// CHECK-DAG: [[TMP35:%.*]] = add nuw i64 [[TMP34]], 4
+// CHECK-DAG: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 4
+// CHECK-DAG: store i64 [[TMP35]], ptr [[TMP38:%[^,]+]],
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK-DAG: [[TMP39:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK-DAG: store ptr @[[RED_INIT2:.+]], ptr [[TMP39]],
@@ -127,7 +126,7 @@ sum = 0.0;
// CHECK-DAG: store ptr [[VLA]], ptr [[TMP52:%[^,]+]],
// CHECK-DAG: [[TMP52]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 1
// CHECK-DAG: [[TMP54:%.*]] = mul nuw i64 [[TMP2]], 4
-// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], ptrtoint (ptr getelementptr (float, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], 4
// CHECK-DAG: store i64 [[TMP54]], ptr [[TMP56:%[^,]+]],
// CHECK-DAG: [[TMP56]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 2
// CHECK-DAG: [[TMP57:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 3
diff --git a/clang/test/OpenMP/master_taskloop_simd_in_reduction_codegen.cpp b/clang/test/OpenMP/master_taskloop_simd_in_reduction_codegen.cpp
index 5fb788dd45e32..0bbed3bbeb7cf 100644
--- a/clang/test/OpenMP/master_taskloop_simd_in_reduction_codegen.cpp
+++ b/clang/test/OpenMP/master_taskloop_simd_in_reduction_codegen.cpp
@@ -145,7 +145,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP35:%.*]] = mul nuw i64 [[TMP2]], 2
-// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 2
// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP37]], align 8
// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
@@ -359,7 +359,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: call void @_ZN1SC1ERKS_(ptr nonnull align 4 dereferenceable(4) [[AGG_RESULT]], ptr nonnull align 4 dereferenceable(4) [[TMP0]])
// CHECK1-NEXT: ret void
//
@@ -451,11 +451,11 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[D]], ptr [[D_ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_1]], ptr [[DOTTASK_RED__ADDR2]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META3]], !align [[META6:![0-9]+]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META3]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META2]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META2]], !align [[META5]]
// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK1-NEXT: [[TMP7:%.*]] = call i32 @__kmpc_master(ptr @[[GLOB1]], i32 [[TMP6]])
@@ -560,66 +560,65 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 8
// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 9
// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[TMP17]], align 8
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META14:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META16:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META15:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP19]], i32 0, i32 1
// CHECK1-NEXT: [[TMP21:%.*]] = load i64, ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
-// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]], !inline_history [[META18:![0-9]+]]
+// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: [[TMP29:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP27]], ptr [[TMP26]])
// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP19]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META3]], !align [[META5]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META2]], !align [[META4]]
// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP21]], 2
-// CHECK1-NEXT: [[TMP33:%.*]] = udiv exact i64 [[TMP32]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
-// CHECK1-NEXT: store i64 [[TMP33]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP35]], ptr [[TMP31]])
-// CHECK1-NEXT: [[TMP37:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP37]] to i32
-// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
+// CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP34]], ptr [[TMP31]])
+// CHECK1-NEXT: [[TMP36:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP36]] to i32
+// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1: omp.inner.for.cond.i:
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19:![0-9]+]]
-// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP39]]
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19:![0-9]+]]
+// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP38]]
// CHECK1-NEXT: br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__EXIT:%.*]]
// CHECK1: omp.inner.for.body.i:
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: store i32 [[TMP40]], ptr [[I_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP41]] to i64
-// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP36]], i64 [[IDXPROM_I]]
-// CHECK1-NEXT: [[TMP42:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP42]] to i32
-// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP43]], [[CONV3_I]]
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: store i32 [[TMP39]], ptr [[I_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP40]] to i64
+// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP35]], i64 [[IDXPROM_I]]
+// CHECK1-NEXT: [[TMP41:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2, !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP41]] to i32
+// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP42]], [[CONV3_I]]
// CHECK1-NEXT: store i32 [[ADD4_I]], ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP44]], 1
-// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP43]], 1
+// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I]], !llvm.loop [[LOOP20:![0-9]+]]
// CHECK1: .omp_outlined..exit:
// CHECK1-NEXT: ret i32 0
@@ -713,32 +712,32 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: store i32 [[CONV]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3: omp.inner.for.cond:
-// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2:![0-9]+]]
+// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1:![0-9]+]]
// CHECK3-NEXT: [[CONV2:%.*]] = sext i32 [[TMP4]] to i64
-// CHECK3-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[CMP:%.*]] = icmp ule i64 [[CONV2]], [[TMP5]]
// CHECK3-NEXT: br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK3: omp.inner.for.body:
-// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP6]], 1
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
-// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP2]]
-// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP1]]
+// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP7]] to i64
// CHECK3-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i16, ptr [[VLA]], i64 [[IDXPROM]]
-// CHECK3-NEXT: [[TMP8:%.*]] = load i16, ptr [[ARRAYIDX]], align 2, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP8:%.*]] = load i16, ptr [[ARRAYIDX]], align 2, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[CONV3:%.*]] = sext i16 [[TMP8]] to i32
-// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP9]], [[CONV3]]
-// CHECK3-NEXT: store i32 [[ADD4]], ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: store i32 [[ADD4]], ptr [[A]], align 4, !llvm.access.group [[ACC_GRP1]]
// 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: [[TMP10:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP10]], 1
-// CHECK3-NEXT: store i32 [[ADD5]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
-// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
+// CHECK3-NEXT: store i32 [[ADD5]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
+// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP2:![0-9]+]]
// CHECK3: omp.inner.for.end:
// CHECK3-NEXT: store i32 5, ptr [[I]], align 4
// CHECK3-NEXT: store i32 0, ptr [[RETVAL]], align 4
diff --git a/clang/test/OpenMP/master_taskloop_simd_reduction_codegen.cpp b/clang/test/OpenMP/master_taskloop_simd_reduction_codegen.cpp
index 676a6c63745b6..164820b0ce124 100644
--- a/clang/test/OpenMP/master_taskloop_simd_reduction_codegen.cpp
+++ b/clang/test/OpenMP/master_taskloop_simd_reduction_codegen.cpp
@@ -90,10 +90,9 @@ sum = 0.0;
// CHECK-DAG: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK-DAG: [[TMP33:%.*]] = ptrtoaddr ptr [[ARRAYIDX5]] to i64
// CHECK-DAG: [[TMP34:%.*]] = sub i64 [[TMP32]], [[TMP33]]
-// CHECK-DAG: [[TMP35:%.*]] = sdiv exact i64 [[TMP34]], 4
-// CHECK-DAG: [[TMP36:%.*]] = add nuw i64 [[TMP35]], 1
-// CHECK-DAG: [[TMP37:%.*]] = mul nuw i64 [[TMP36]], ptrtoint (ptr getelementptr (%struct.S, ptr null, i32 1) to i64)
-// CHECK-DAG: store i64 [[TMP37]], ptr [[TMP38:%[^,]+]],
+// CHECK-DAG: [[TMP35:%.*]] = add nuw i64 [[TMP34]], 4
+// CHECK-DAG: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 4
+// CHECK-DAG: store i64 [[TMP35]], ptr [[TMP38:%[^,]+]],
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK-DAG: [[TMP39:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK-DAG: store ptr @[[RED_INIT2:.+]], ptr [[TMP39]],
@@ -123,7 +122,7 @@ sum = 0.0;
// CHECK-DAG: store ptr [[VLA]], ptr [[TMP52:%[^,]+]],
// CHECK-DAG: [[TMP52]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8:%.+]], i32 0, i32 1
// CHECK-DAG: [[TMP54:%.*]] = mul nuw i64 [[TMP2]], 4
-// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], ptrtoint (ptr getelementptr (float, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], 4
// CHECK-DAG: store i64 [[TMP54]], ptr [[TMP56:%[^,]+]],
// CHECK-DAG: [[TMP56]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 2
// CHECK-DAG: [[TMP57:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 3
diff --git a/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
index 0849dad839de5..10490e3bf42bd 100644
--- a/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
@@ -76,7 +76,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: store i64 0, ptr [[DOTOMP_LB]], align 8
// CHECK1-NEXT: store i64 9, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: store i64 1, ptr [[DOTOMP_STRIDE]], align 8
@@ -97,223 +97,221 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP6]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP6]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP6]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_7:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
-// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
-// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN10]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP49]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
+// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
+// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN10]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP47]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP48:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP48]], 9
// 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: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP49]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP50]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP51]], [[TMP52]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP55]], 1
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP53]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP56]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP57]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP59]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
-// CHECK1-NEXT: [[TMP62:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP61]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP62]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP63]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[TMP64]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP65]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP62]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP66]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP68]], ptr [[TMP67]], align 8
-// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[TMP69]], align 4
-// CHECK1-NEXT: [[TMP71:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP70]], ptr [[TMP62]])
+// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP54]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP55]], align 8
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP57:%.*]] = load ptr, ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP57]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP58:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load i32, ptr [[TMP58]], align 4
+// CHECK1-NEXT: [[TMP60:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP59]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP61:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP60]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP61]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP63:%.*]] = load ptr, ptr [[TMP62]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP63]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP60]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP64]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP66]], ptr [[TMP65]], align 8
+// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP68:%.*]] = load i32, ptr [[TMP67]], align 4
+// CHECK1-NEXT: [[TMP69:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP68]], ptr [[TMP60]])
// 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: [[TMP72:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP72]], 1
+// CHECK1-NEXT: [[TMP70:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP70]], 1
// CHECK1-NEXT: store i64 [[ADD14]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP71:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[TMP71]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP72]])
// CHECK1-NEXT: [[TMP73:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[TMP73]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP74]])
-// CHECK1-NEXT: [[TMP75:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[TMP75]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP76]], i32 1)
-// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP77]], align 8
-// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP78]], align 8
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP80:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP80]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP81:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP81]], align 4
-// CHECK1-NEXT: [[TMP83:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3:[0-9]+]], i32 [[TMP82]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP83]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP74]], i32 1)
+// CHECK1-NEXT: [[TMP75:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP75]], align 8
+// CHECK1-NEXT: [[TMP76:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP76]], align 8
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP78:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP78]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP79:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP80:%.*]] = load i32, ptr [[TMP79]], align 4
+// CHECK1-NEXT: [[TMP81:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3:[0-9]+]], i32 [[TMP80]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP81]], 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: [[TMP84:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP85:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP84]], [[TMP85]]
+// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP83:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP82]], [[TMP83]]
// CHECK1-NEXT: store i32 [[ADD15]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP86:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP86]]
+// CHECK1-NEXT: [[TMP84:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP84]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE22:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST16:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT20:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP87]] to i32
-// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP88]] to i32
+// CHECK1-NEXT: [[TMP85:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP85]] to i32
+// CHECK1-NEXT: [[TMP86:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP86]] to i32
// CHECK1-NEXT: [[ADD18:%.*]] = add nsw i32 [[CONV]], [[CONV17]]
// CHECK1-NEXT: [[CONV19:%.*]] = trunc i32 [[ADD18]] to i8
// CHECK1-NEXT: store i8 [[CONV19]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT20]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP86]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP84]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_DONE22]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done22:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP80]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP89:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP90:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP89]] monotonic, align 4
-// CHECK1-NEXT: [[TMP91:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP91]]
+// CHECK1-NEXT: [[TMP87:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP88:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP87]] monotonic, align 4
+// CHECK1-NEXT: [[TMP89:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP89]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY23]], label [[OMP_ARRAYCPY_DONE36:%.*]], label [[OMP_ARRAYCPY_BODY24:%.*]]
// CHECK1: omp.arraycpy.body24:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST25:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT34:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT33:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP92]] to i32
+// CHECK1-NEXT: [[TMP90:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP90]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP93:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP98:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP93]], ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP94]] to i32
-// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP95]] to i32
+// CHECK1-NEXT: [[TMP91:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP96:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP91]], ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP92]] to i32
+// CHECK1-NEXT: [[TMP93:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP93]] to i32
// CHECK1-NEXT: [[ADD31:%.*]] = add nsw i32 [[CONV29]], [[CONV30]]
// CHECK1-NEXT: [[CONV32:%.*]] = trunc i32 [[ADD31]] to i8
// CHECK1-NEXT: store i8 [[CONV32]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP97:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP93]], i8 [[TMP96]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP98]] = extractvalue { i8, i1 } [[TMP97]], 0
-// CHECK1-NEXT: [[TMP99:%.*]] = extractvalue { i8, i1 } [[TMP97]], 1
-// CHECK1-NEXT: br i1 [[TMP99]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP95:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP91]], i8 [[TMP94]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP96]] = extractvalue { i8, i1 } [[TMP95]], 0
+// CHECK1-NEXT: [[TMP97:%.*]] = extractvalue { i8, i1 } [[TMP95]], 1
+// CHECK1-NEXT: br i1 [[TMP97]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT34]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP91]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP89]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_DONE36]], label [[OMP_ARRAYCPY_BODY24]]
// CHECK1: omp.arraycpy.done36:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP100]])
+// CHECK1-NEXT: [[TMP98:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP98]])
// CHECK1-NEXT: ret void
//
//
@@ -437,31 +435,31 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]]
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META4:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]], !inline_history [[META14:![0-9]+]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13]]
// CHECK1-NEXT: [[TMP17:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP15]], ptr [[TMP14]])
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[TMP18]], align 8
// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[TMP22]], align 4
// CHECK1-NEXT: [[TMP24:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[LB_ADD_LEN_I:%.*]] = add nsw i64 -1, [[TMP24]]
@@ -474,19 +472,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
-// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
+// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
index 5428cd551adc8..32165cb557fa1 100644
--- a/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
@@ -69,7 +69,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[ARGC1]], align 4
// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP1]], i64 0
@@ -86,189 +86,187 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 0, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: [[TMP50:%.*]] = call i32 @__kmpc_master(ptr @[[GLOB1]], i32 [[TMP49]])
-// CHECK1-NEXT: [[TMP51:%.*]] = icmp ne i32 [[TMP50]], 0
-// CHECK1-NEXT: br i1 [[TMP51]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 0, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: [[TMP48:%.*]] = call i32 @__kmpc_master(ptr @[[GLOB1]], i32 [[TMP47]])
+// CHECK1-NEXT: [[TMP49:%.*]] = icmp ne i32 [[TMP48]], 0
+// CHECK1-NEXT: br i1 [[TMP49]], label [[OMP_IF_THEN:%.*]], label [[OMP_IF_END:%.*]]
// CHECK1: omp_if.then:
-// CHECK1-NEXT: [[TMP52:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP52]], align 8
-// CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP53]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP55:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP55]], ptr [[TMP54]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = load i32, ptr [[TMP56]], align 4
-// CHECK1-NEXT: [[TMP58:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP57]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP58]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP59]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[TMP60]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP61]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP58]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP62]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP64:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP64]], ptr [[TMP63]], align 8
-// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP66:%.*]] = load i32, ptr [[TMP65]], align 4
-// CHECK1-NEXT: [[TMP67:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP66]], ptr [[TMP58]])
-// CHECK1-NEXT: call void @__kmpc_end_master(ptr @[[GLOB1]], i32 [[TMP49]])
+// CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP50]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP51]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP53:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: store ptr [[TMP53]], ptr [[TMP52]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = load i32, ptr [[TMP54]], align 4
+// CHECK1-NEXT: [[TMP56:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP55]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP56]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP57]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[TMP58]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP59]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP56]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP61:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP60]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP62]], ptr [[TMP61]], align 8
+// CHECK1-NEXT: [[TMP63:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP64:%.*]] = load i32, ptr [[TMP63]], align 4
+// CHECK1-NEXT: [[TMP65:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP64]], ptr [[TMP56]])
+// CHECK1-NEXT: call void @__kmpc_end_master(ptr @[[GLOB1]], i32 [[TMP47]])
// CHECK1-NEXT: br label [[OMP_IF_END]]
// CHECK1: omp_if.end:
-// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP69]], i32 0)
-// CHECK1-NEXT: [[TMP70:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP70]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP71]], align 8
-// CHECK1-NEXT: [[TMP72:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP73:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP73]], ptr [[TMP72]], align 8
-// CHECK1-NEXT: [[TMP74:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP75:%.*]] = load i32, ptr [[TMP74]], align 4
-// CHECK1-NEXT: [[TMP76:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP75]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP76]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP67:%.*]] = load i32, ptr [[TMP66]], align 4
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP67]], i32 0)
+// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP68]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP69]], align 8
+// CHECK1-NEXT: [[TMP70:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP71:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP71]], ptr [[TMP70]], align 8
+// CHECK1-NEXT: [[TMP72:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP73:%.*]] = load i32, ptr [[TMP72]], align 4
+// CHECK1-NEXT: [[TMP74:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP73]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP74]], 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: [[TMP77:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP77]], [[TMP78]]
+// CHECK1-NEXT: [[TMP75:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP75]], [[TMP76]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP79]]
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP77]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE18:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST12:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT16:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP80:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP80]] to i32
-// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP81]] to i32
+// CHECK1-NEXT: [[TMP78:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP78]] to i32
+// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP79]] to i32
// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[CONV]], [[CONV13]]
// CHECK1-NEXT: [[CONV15:%.*]] = trunc i32 [[ADD14]] to i8
// CHECK1-NEXT: store i8 [[CONV15]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT16]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP79]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP77]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE17]], label [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done18:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP75]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP73]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP83:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP82]] monotonic, align 4
-// CHECK1-NEXT: [[TMP84:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP84]]
+// CHECK1-NEXT: [[TMP80:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP81:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP80]] monotonic, align 4
+// CHECK1-NEXT: [[TMP82:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP82]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY19]], label [[OMP_ARRAYCPY_DONE32:%.*]], label [[OMP_ARRAYCPY_BODY20:%.*]]
// CHECK1: omp.arraycpy.body20:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST21:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT30:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST22:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT29:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP85:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP85]] to i32
+// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP83]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP86:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP91:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP86]], ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP87]] to i32
-// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP88]] to i32
+// CHECK1-NEXT: [[TMP84:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP89:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP84]], ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[TMP85:%.*]] = load i8, ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP85]] to i32
+// CHECK1-NEXT: [[TMP86:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP86]] to i32
// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[CONV25]], [[CONV26]]
// CHECK1-NEXT: [[CONV28:%.*]] = trunc i32 [[ADD27]] to i8
// CHECK1-NEXT: store i8 [[CONV28]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP90:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP86]], i8 [[TMP89]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP91]] = extractvalue { i8, i1 } [[TMP90]], 0
-// CHECK1-NEXT: [[TMP92:%.*]] = extractvalue { i8, i1 } [[TMP90]], 1
-// CHECK1-NEXT: br i1 [[TMP92]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP88:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP84]], i8 [[TMP87]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP89]] = extractvalue { i8, i1 } [[TMP88]], 0
+// CHECK1-NEXT: [[TMP90:%.*]] = extractvalue { i8, i1 } [[TMP88]], 1
+// CHECK1-NEXT: br i1 [[TMP90]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT29]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT30]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP84]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP82]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE31]], label [[OMP_ARRAYCPY_DONE32]], label [[OMP_ARRAYCPY_BODY20]]
// CHECK1: omp.arraycpy.done32:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP93:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP93]])
+// CHECK1-NEXT: [[TMP91:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP91]])
// CHECK1-NEXT: ret void
//
//
@@ -392,31 +390,31 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]]
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META4:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]], !inline_history [[META14:![0-9]+]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13]]
// CHECK1-NEXT: [[TMP17:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP15]], ptr [[TMP14]])
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[TMP18]], align 8
// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[TMP22]], align 4
// CHECK1-NEXT: [[TMP24:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[LB_ADD_LEN_I:%.*]] = add nsw i64 -1, [[TMP24]]
@@ -429,19 +427,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
-// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
+// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_master_taskloop_reduction_codegen.cpp b/clang/test/OpenMP/parallel_master_taskloop_reduction_codegen.cpp
index 01bf4519e012b..583d694e876de 100644
--- a/clang/test/OpenMP/parallel_master_taskloop_reduction_codegen.cpp
+++ b/clang/test/OpenMP/parallel_master_taskloop_reduction_codegen.cpp
@@ -94,10 +94,9 @@ sum = 0.0;
// CHECK-DAG: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK-DAG: [[TMP33:%.*]] = ptrtoaddr ptr [[ARRAYIDX5]] to i64
// CHECK-DAG: [[TMP34:%.*]] = sub i64 [[TMP32]], [[TMP33]]
-// CHECK-DAG: [[TMP35:%.*]] = sdiv exact i64 [[TMP34]], 4
-// CHECK-DAG: [[TMP36:%.*]] = add nuw i64 [[TMP35]], 1
-// CHECK-DAG: [[TMP37:%.*]] = mul nuw i64 [[TMP36]], ptrtoint (ptr getelementptr (%struct.S, ptr null, i32 1) to i64)
-// CHECK-DAG: store i64 [[TMP37]], ptr [[TMP38:%[^,]+]],
+// CHECK-DAG: [[TMP35:%.*]] = add nuw i64 [[TMP34]], 4
+// CHECK-DAG: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 4
+// CHECK-DAG: store i64 [[TMP35]], ptr [[TMP38:%[^,]+]],
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK-DAG: [[TMP39:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK-DAG: store ptr @[[RED_INIT2:.+]], ptr [[TMP39]],
@@ -127,7 +126,7 @@ sum = 0.0;
// CHECK-DAG: store ptr [[VLA:%.+]], ptr [[TMP52:%[^,]+]],
// CHECK-DAG: [[TMP52]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8:%.+]], i32 0, i32 1
// CHECK-DAG: [[TMP54:%.*]] = mul nuw i64 [[TMP2:%.+]], 4
-// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], ptrtoint (ptr getelementptr (float, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], 4
// CHECK-DAG: store i64 [[TMP54]], ptr [[TMP56:%[^,]+]],
// CHECK-DAG: [[TMP56]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 2
// CHECK-DAG: [[TMP57:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 3
diff --git a/clang/test/OpenMP/parallel_master_taskloop_simd_reduction_codegen.cpp b/clang/test/OpenMP/parallel_master_taskloop_simd_reduction_codegen.cpp
index dfbf84a186da7..4abcf10df552e 100644
--- a/clang/test/OpenMP/parallel_master_taskloop_simd_reduction_codegen.cpp
+++ b/clang/test/OpenMP/parallel_master_taskloop_simd_reduction_codegen.cpp
@@ -94,10 +94,9 @@ sum = 0.0;
// CHECK-DAG: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK-DAG: [[TMP33:%.*]] = ptrtoaddr ptr [[ARRAYIDX5]] to i64
// CHECK-DAG: [[TMP34:%.*]] = sub i64 [[TMP32]], [[TMP33]]
-// CHECK-DAG: [[TMP35:%.*]] = sdiv exact i64 [[TMP34]], 4
-// CHECK-DAG: [[TMP36:%.*]] = add nuw i64 [[TMP35]], 1
-// CHECK-DAG: [[TMP37:%.*]] = mul nuw i64 [[TMP36]], ptrtoint (ptr getelementptr (%struct.S, ptr null, i32 1) to i64)
-// CHECK-DAG: store i64 [[TMP37]], ptr [[TMP38:%[^,]+]],
+// CHECK-DAG: [[TMP35:%.*]] = add nuw i64 [[TMP34]], 4
+// CHECK-DAG: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 4
+// CHECK-DAG: store i64 [[TMP35]], ptr [[TMP38:%[^,]+]],
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK-DAG: [[TMP39:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK-DAG: store ptr @[[RED_INIT2:.+]], ptr [[TMP39]],
@@ -127,7 +126,7 @@ sum = 0.0;
// CHECK-DAG: store ptr [[VLA:%.+]], ptr [[TMP52:%[^,]+]],
// CHECK-DAG: [[TMP52]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 1
// CHECK-DAG: [[TMP54:%.*]] = mul nuw i64 [[TMP2:%.+]], 4
-// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], ptrtoint (ptr getelementptr (float, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], 4
// CHECK-DAG: store i64 [[TMP54]], ptr [[TMP56:%[^,]+]],
// CHECK-DAG: [[TMP56]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 2
// CHECK-DAG: [[TMP57:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 3
diff --git a/clang/test/OpenMP/parallel_reduction_codegen.cpp b/clang/test/OpenMP/parallel_reduction_codegen.cpp
index 4bec2dbc8a209..beef3b29f5029 100644
--- a/clang/test/OpenMP/parallel_reduction_codegen.cpp
+++ b/clang/test/OpenMP/parallel_reduction_codegen.cpp
@@ -360,99 +360,98 @@ int main() {
// CHECK1-NEXT: [[TMP2:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
// CHECK1-NEXT: [[TMP3:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK1-NEXT: [[TMP4:%.*]] = sub i64 [[TMP2]], [[TMP3]]
-// CHECK1-NEXT: [[TMP5:%.*]] = sdiv exact i64 [[TMP4]], 2
-// CHECK1-NEXT: [[TMP6:%.*]] = add nuw i64 [[TMP5]], 1
-// CHECK1-NEXT: [[TMP7:%.*]] = mul nuw i64 [[TMP6]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP8:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP5:%.*]] = add nuw i64 [[TMP4]], 2
+// CHECK1-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], 2
+// CHECK1-NEXT: [[TMP7:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i16, i64 [[TMP6]], align 16
// CHECK1-NEXT: store i64 [[TMP6]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP9:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP6]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP9]]
+// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP6]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP8]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i16 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP9]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP8]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[X_ADDR]], align 8
-// CHECK1-NEXT: [[TMP11:%.*]] = ptrtoaddr ptr [[TMP10]] to i64
-// CHECK1-NEXT: [[TMP12:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
-// CHECK1-NEXT: [[TMP13:%.*]] = sub i64 [[TMP11]], [[TMP12]]
-// CHECK1-NEXT: [[TMP14:%.*]] = sdiv exact i64 [[TMP13]], 2
-// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP14]]
-// CHECK1-NEXT: store ptr [[TMP15]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP16]], align 8
-// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP18:%.*]] = inttoptr i64 [[TMP6]] to ptr
-// CHECK1-NEXT: store ptr [[TMP18]], ptr [[TMP17]], align 8
-// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP20:%.*]] = load i32, ptr [[TMP19]], align 4
-// CHECK1-NEXT: [[TMP21:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1:[0-9]+]], i32 [[TMP20]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_Z14foo_array_sectPs.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP21]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[X_ADDR]], align 8
+// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[TMP9]] to i64
+// CHECK1-NEXT: [[TMP11:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
+// CHECK1-NEXT: [[TMP12:%.*]] = sub i64 [[TMP10]], [[TMP11]]
+// CHECK1-NEXT: [[TMP13:%.*]] = sdiv exact i64 [[TMP12]], 2
+// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP13]]
+// CHECK1-NEXT: store ptr [[TMP14]], ptr [[TMP]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP15]], align 8
+// CHECK1-NEXT: [[TMP16:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP17:%.*]] = inttoptr i64 [[TMP6]] to ptr
+// CHECK1-NEXT: store ptr [[TMP17]], ptr [[TMP16]], align 8
+// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP19:%.*]] = load i32, ptr [[TMP18]], align 4
+// CHECK1-NEXT: [[TMP20:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1:[0-9]+]], i32 [[TMP19]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_Z14foo_array_sectPs.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP20]], 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: [[TMP22:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP22]]
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP21]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE7:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST2:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT5:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP23:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
-// CHECK1-NEXT: [[CONV:%.*]] = sext i16 [[TMP23]] to i32
-// CHECK1-NEXT: [[TMP24:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
-// CHECK1-NEXT: [[CONV3:%.*]] = sext i16 [[TMP24]] to i32
+// CHECK1-NEXT: [[TMP22:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
+// CHECK1-NEXT: [[CONV:%.*]] = sext i16 [[TMP22]] to i32
+// CHECK1-NEXT: [[TMP23:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
+// CHECK1-NEXT: [[CONV3:%.*]] = sext i16 [[TMP23]] to i32
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[CONV]], [[CONV3]]
// CHECK1-NEXT: [[CONV4:%.*]] = trunc i32 [[ADD]] to i16
// CHECK1-NEXT: store i16 [[CONV4]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT5]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE6:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP22]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE6:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP21]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE6]], label [[OMP_ARRAYCPY_DONE7]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done7:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP20]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP19]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY8:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP25]]
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY8:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP24]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY8]], label [[OMP_ARRAYCPY_DONE21:%.*]], label [[OMP_ARRAYCPY_BODY9:%.*]]
// CHECK1: omp.arraycpy.body9:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST10:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT19:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST11:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT18:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP26:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
-// CHECK1-NEXT: [[CONV12:%.*]] = sext i16 [[TMP26]] to i32
+// CHECK1-NEXT: [[TMP25:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
+// CHECK1-NEXT: [[CONV12:%.*]] = sext i16 [[TMP25]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]] monotonic, align 2
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP27:%.*]] = phi i16 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY9]] ], [ [[TMP32:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i16 [[TMP27]], ptr [[_TMP13]], align 2
-// CHECK1-NEXT: [[TMP28:%.*]] = load i16, ptr [[_TMP13]], align 2
-// CHECK1-NEXT: [[CONV14:%.*]] = sext i16 [[TMP28]] to i32
-// CHECK1-NEXT: [[TMP29:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
-// CHECK1-NEXT: [[CONV15:%.*]] = sext i16 [[TMP29]] to i32
+// CHECK1-NEXT: [[TMP26:%.*]] = phi i16 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY9]] ], [ [[TMP31:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i16 [[TMP26]], ptr [[_TMP13]], align 2
+// CHECK1-NEXT: [[TMP27:%.*]] = load i16, ptr [[_TMP13]], align 2
+// CHECK1-NEXT: [[CONV14:%.*]] = sext i16 [[TMP27]] to i32
+// CHECK1-NEXT: [[TMP28:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
+// CHECK1-NEXT: [[CONV15:%.*]] = sext i16 [[TMP28]] to i32
// CHECK1-NEXT: [[ADD16:%.*]] = add nsw i32 [[CONV14]], [[CONV15]]
// CHECK1-NEXT: [[CONV17:%.*]] = trunc i32 [[ADD16]] to i16
// CHECK1-NEXT: store i16 [[CONV17]], ptr [[ATOMIC_TEMP]], align 2
-// CHECK1-NEXT: [[TMP30:%.*]] = load i16, ptr [[ATOMIC_TEMP]], align 2
-// CHECK1-NEXT: [[TMP31:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i16 [[TMP27]], i16 [[TMP30]] monotonic monotonic, align 2
-// CHECK1-NEXT: [[TMP32]] = extractvalue { i16, i1 } [[TMP31]], 0
-// CHECK1-NEXT: [[TMP33:%.*]] = extractvalue { i16, i1 } [[TMP31]], 1
-// CHECK1-NEXT: br i1 [[TMP33]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP29:%.*]] = load i16, ptr [[ATOMIC_TEMP]], align 2
+// CHECK1-NEXT: [[TMP30:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i16 [[TMP26]], i16 [[TMP29]] monotonic monotonic, align 2
+// CHECK1-NEXT: [[TMP31]] = extractvalue { i16, i1 } [[TMP30]], 0
+// CHECK1-NEXT: [[TMP32:%.*]] = extractvalue { i16, i1 } [[TMP30]], 1
+// CHECK1-NEXT: br i1 [[TMP32]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT18]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT19]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT18]], [[TMP25]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT18]], [[TMP24]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_BODY9]]
// CHECK1: omp.arraycpy.done21:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP34]])
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP33]])
// CHECK1-NEXT: ret void
//
//
@@ -608,12 +607,12 @@ int main() {
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR1]], ptr [[VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5:![0-9]+]], !align [[META6:![0-9]+]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4:![0-9]+]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store float 0.000000e+00, ptr [[T_VAR2]], align 4
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR14]])
@@ -843,12 +842,12 @@ int main() {
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR1]], ptr [[VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store float 0.000000e+00, ptr [[T_VAR2]], align 4
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT: call void @_ZN1SIfEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR14]])
@@ -863,7 +862,7 @@ int main() {
// CHECK1-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4
// CHECK1-NEXT: [[ARRAYIDX6:%.*]] = getelementptr inbounds [2 x [[STRUCT_S]]], ptr [[TMP2]], i64 0, i64 0
// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 4 [[ARRAYIDX6]], ptr align 4 [[VAR3]], i64 4, i1 false)
-// CHECK1-NEXT: br label [[WHILE_COND]], !llvm.loop [[LOOP7:![0-9]+]]
+// CHECK1-NEXT: br label [[WHILE_COND]], !llvm.loop [[LOOP6:![0-9]+]]
//
//
// CHECK1-LABEL: define {{[^@]+}}@main.omp_outlined.2
@@ -880,7 +879,7 @@ int main() {
// 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 [[CF]], ptr [[CF_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[CF_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[CF_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[CF1_REALP:%.*]] = getelementptr inbounds nuw { float, float }, ptr [[CF1]], i32 0, i32 0
// CHECK1-NEXT: [[CF1_IMAGP:%.*]] = getelementptr inbounds nuw { float, float }, ptr [[CF1]], i32 0, i32 1
// CHECK1-NEXT: store float 0.000000e+00, ptr [[CF1_REALP]], align 4
@@ -1039,15 +1038,15 @@ int main() {
// CHECK1-NEXT: [[BF_SET:%.*]] = or i8 [[BF_CLEAR]], 0
// CHECK1-NEXT: store i8 [[BF_SET]], ptr [[B]], align 4
// CHECK1-NEXT: [[C:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[C]], align 8
// CHECK1-NEXT: [[A3:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 0
// CHECK1-NEXT: store ptr [[A3]], ptr [[A2]], align 8
// CHECK1-NEXT: [[C6:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[C6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[C6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[C5]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 4, ptr @_ZN2SSC2ERi.omp_outlined, ptr [[THIS1]], ptr [[TMP2]], ptr [[B4]], ptr [[TMP3]])
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[B4]], align 4
// CHECK1-NEXT: [[B7:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 1
@@ -1084,26 +1083,26 @@ int main() {
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK1-NEXT: store ptr [[C]], ptr [[C_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
// CHECK1-NEXT: store ptr [[TMP3]], ptr [[_TMP1]], align 8
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[A2]], align 4
// CHECK1-NEXT: store ptr [[A2]], ptr [[_TMP3]], align 8
// CHECK1-NEXT: store i32 0, ptr [[B4]], align 4
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[C5]], align 4
// CHECK1-NEXT: store ptr [[C5]], ptr [[_TMP6]], align 8
-// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP7:%.*]] = load i32, ptr [[TMP6]], align 4
// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[TMP6]], align 4
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, ptr [[B4]], align 4
// CHECK1-NEXT: [[DEC:%.*]] = add nsw i32 [[TMP8]], -1
// CHECK1-NEXT: store i32 [[DEC]], ptr [[B4]], align 4
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
// CHECK1-NEXT: [[DIV:%.*]] = sdiv i32 [[TMP10]], 1
// CHECK1-NEXT: store i32 [[DIV]], ptr [[TMP9]], align 4
@@ -1281,12 +1280,12 @@ int main() {
// CHECK1-NEXT: store ptr [[VAR]], ptr [[VAR_ADDR]], align 8
// CHECK1-NEXT: store ptr [[VAR1]], ptr [[VAR1_ADDR]], align 8
// CHECK1-NEXT: store ptr [[T_VAR1]], ptr [[T_VAR1_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[VEC_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[T_VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[S_ARR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[VAR_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK1-NEXT: [[TMP5:%.*]] = load ptr, ptr [[T_VAR1_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 0, ptr [[T_VAR2]], align 128
// CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR3]])
// CHECK1-NEXT: call void @_ZN1SIiEC1Ev(ptr noundef nonnull align 4 dereferenceable(4) [[VAR14]])
@@ -1491,7 +1490,7 @@ int main() {
// CHECK1-NEXT: store i32 0, ptr [[A]], align 4
// CHECK1-NEXT: [[A3:%.*]] = getelementptr inbounds nuw [[STRUCT_SST]], ptr [[THIS1]], i32 0, i32 0
// CHECK1-NEXT: store ptr [[A3]], ptr [[A2]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 2, ptr @_ZN3SSTIiEC2Ev.omp_outlined, ptr [[THIS1]], ptr [[TMP0]])
// CHECK1-NEXT: ret void
//
@@ -1514,12 +1513,12 @@ int main() {
// CHECK1-NEXT: store ptr [[THIS]], ptr [[THIS_ADDR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: store i32 1, ptr [[A1]], align 4
// CHECK1-NEXT: store ptr [[A1]], ptr [[_TMP2]], align 8
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP2]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[_TMP2]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, ptr [[TMP3]], align 4
// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[TMP3]], align 4
@@ -1638,99 +1637,98 @@ int main() {
// CHECK3-NEXT: [[TMP2:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
// CHECK3-NEXT: [[TMP3:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK3-NEXT: [[TMP4:%.*]] = sub i64 [[TMP2]], [[TMP3]]
-// CHECK3-NEXT: [[TMP5:%.*]] = sdiv exact i64 [[TMP4]], 2
-// CHECK3-NEXT: [[TMP6:%.*]] = add nuw i64 [[TMP5]], 1
-// CHECK3-NEXT: [[TMP7:%.*]] = mul nuw i64 [[TMP6]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK3-NEXT: [[TMP8:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK3-NEXT: store ptr [[TMP8]], ptr [[SAVED_STACK]], align 8
+// CHECK3-NEXT: [[TMP5:%.*]] = add nuw i64 [[TMP4]], 2
+// CHECK3-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], 2
+// CHECK3-NEXT: [[TMP7:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK3-NEXT: store ptr [[TMP7]], ptr [[SAVED_STACK]], align 8
// CHECK3-NEXT: [[VLA:%.*]] = alloca i16, i64 [[TMP6]], align 16
// CHECK3-NEXT: store i64 [[TMP6]], ptr [[__VLA_EXPR0]], align 8
-// CHECK3-NEXT: [[TMP9:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP6]]
-// CHECK3-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP9]]
+// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP6]]
+// CHECK3-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP8]]
// CHECK3-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK3: omp.arrayinit.body:
// CHECK3-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK3-NEXT: store i16 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK3-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK3-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP9]]
+// CHECK3-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP8]]
// CHECK3-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK3: omp.arrayinit.done:
-// CHECK3-NEXT: [[TMP10:%.*]] = load ptr, ptr [[X_ADDR]], align 8
-// CHECK3-NEXT: [[TMP11:%.*]] = ptrtoaddr ptr [[TMP10]] to i64
-// CHECK3-NEXT: [[TMP12:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
-// CHECK3-NEXT: [[TMP13:%.*]] = sub i64 [[TMP11]], [[TMP12]]
-// CHECK3-NEXT: [[TMP14:%.*]] = sdiv exact i64 [[TMP13]], 2
-// CHECK3-NEXT: [[TMP15:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP14]]
-// CHECK3-NEXT: store ptr [[TMP15]], ptr [[TMP]], align 8
-// CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK3-NEXT: store ptr [[VLA]], ptr [[TMP16]], align 8
-// CHECK3-NEXT: [[TMP17:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK3-NEXT: [[TMP18:%.*]] = inttoptr i64 [[TMP6]] to ptr
-// CHECK3-NEXT: store ptr [[TMP18]], ptr [[TMP17]], align 8
-// CHECK3-NEXT: [[TMP19:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK3-NEXT: [[TMP20:%.*]] = load i32, ptr [[TMP19]], align 4
-// CHECK3-NEXT: [[TMP21:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1:[0-9]+]], i32 [[TMP20]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_Z14foo_array_sectPs.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK3-NEXT: switch i32 [[TMP21]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[X_ADDR]], align 8
+// CHECK3-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[TMP9]] to i64
+// CHECK3-NEXT: [[TMP11:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
+// CHECK3-NEXT: [[TMP12:%.*]] = sub i64 [[TMP10]], [[TMP11]]
+// CHECK3-NEXT: [[TMP13:%.*]] = sdiv exact i64 [[TMP12]], 2
+// CHECK3-NEXT: [[TMP14:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP13]]
+// CHECK3-NEXT: store ptr [[TMP14]], ptr [[TMP]], align 8
+// CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK3-NEXT: store ptr [[VLA]], ptr [[TMP15]], align 8
+// CHECK3-NEXT: [[TMP16:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK3-NEXT: [[TMP17:%.*]] = inttoptr i64 [[TMP6]] to ptr
+// CHECK3-NEXT: store ptr [[TMP17]], ptr [[TMP16]], align 8
+// CHECK3-NEXT: [[TMP18:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK3-NEXT: [[TMP19:%.*]] = load i32, ptr [[TMP18]], align 4
+// CHECK3-NEXT: [[TMP20:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1:[0-9]+]], i32 [[TMP19]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_Z14foo_array_sectPs.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK3-NEXT: switch i32 [[TMP20]], 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: [[TMP22:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
-// CHECK3-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP22]]
+// CHECK3-NEXT: [[TMP21:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
+// CHECK3-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP21]]
// CHECK3-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE7:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK3: omp.arraycpy.body:
// CHECK3-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK3-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST2:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT5:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK3-NEXT: [[TMP23:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
-// CHECK3-NEXT: [[CONV:%.*]] = sext i16 [[TMP23]] to i32
-// CHECK3-NEXT: [[TMP24:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
-// CHECK3-NEXT: [[CONV3:%.*]] = sext i16 [[TMP24]] to i32
+// CHECK3-NEXT: [[TMP22:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
+// CHECK3-NEXT: [[CONV:%.*]] = sext i16 [[TMP22]] to i32
+// CHECK3-NEXT: [[TMP23:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
+// CHECK3-NEXT: [[CONV3:%.*]] = sext i16 [[TMP23]] to i32
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 [[CONV]], [[CONV3]]
// CHECK3-NEXT: [[CONV4:%.*]] = trunc i32 [[ADD]] to i16
// CHECK3-NEXT: store i16 [[CONV4]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
// CHECK3-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT5]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], i32 1
// CHECK3-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK3-NEXT: [[OMP_ARRAYCPY_DONE6:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP22]]
+// CHECK3-NEXT: [[OMP_ARRAYCPY_DONE6:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP21]]
// CHECK3-NEXT: br i1 [[OMP_ARRAYCPY_DONE6]], label [[OMP_ARRAYCPY_DONE7]], label [[OMP_ARRAYCPY_BODY]]
// CHECK3: omp.arraycpy.done7:
-// CHECK3-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP20]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK3-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP19]], ptr @.gomp_critical_user_.reduction.var)
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3: .omp.reduction.case2:
-// CHECK3-NEXT: [[TMP25:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
-// CHECK3-NEXT: [[OMP_ARRAYCPY_ISEMPTY8:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP25]]
+// CHECK3-NEXT: [[TMP24:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
+// CHECK3-NEXT: [[OMP_ARRAYCPY_ISEMPTY8:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP24]]
// CHECK3-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY8]], label [[OMP_ARRAYCPY_DONE21:%.*]], label [[OMP_ARRAYCPY_BODY9:%.*]]
// CHECK3: omp.arraycpy.body9:
// CHECK3-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST10:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT19:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK3-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST11:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT18:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK3-NEXT: [[TMP26:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
-// CHECK3-NEXT: [[CONV12:%.*]] = sext i16 [[TMP26]] to i32
+// CHECK3-NEXT: [[TMP25:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
+// CHECK3-NEXT: [[CONV12:%.*]] = sext i16 [[TMP25]] to i32
// CHECK3-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]] monotonic, align 2
// CHECK3-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK3: atomic_cont:
-// CHECK3-NEXT: [[TMP27:%.*]] = phi i16 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY9]] ], [ [[TMP32:%.*]], [[ATOMIC_CONT]] ]
-// CHECK3-NEXT: store i16 [[TMP27]], ptr [[_TMP13]], align 2
-// CHECK3-NEXT: [[TMP28:%.*]] = load i16, ptr [[_TMP13]], align 2
-// CHECK3-NEXT: [[CONV14:%.*]] = sext i16 [[TMP28]] to i32
-// CHECK3-NEXT: [[TMP29:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
-// CHECK3-NEXT: [[CONV15:%.*]] = sext i16 [[TMP29]] to i32
+// CHECK3-NEXT: [[TMP26:%.*]] = phi i16 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY9]] ], [ [[TMP31:%.*]], [[ATOMIC_CONT]] ]
+// CHECK3-NEXT: store i16 [[TMP26]], ptr [[_TMP13]], align 2
+// CHECK3-NEXT: [[TMP27:%.*]] = load i16, ptr [[_TMP13]], align 2
+// CHECK3-NEXT: [[CONV14:%.*]] = sext i16 [[TMP27]] to i32
+// CHECK3-NEXT: [[TMP28:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
+// CHECK3-NEXT: [[CONV15:%.*]] = sext i16 [[TMP28]] to i32
// CHECK3-NEXT: [[ADD16:%.*]] = add nsw i32 [[CONV14]], [[CONV15]]
// CHECK3-NEXT: [[CONV17:%.*]] = trunc i32 [[ADD16]] to i16
// CHECK3-NEXT: store i16 [[CONV17]], ptr [[ATOMIC_TEMP]], align 2
-// CHECK3-NEXT: [[TMP30:%.*]] = load i16, ptr [[ATOMIC_TEMP]], align 2
-// CHECK3-NEXT: [[TMP31:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i16 [[TMP27]], i16 [[TMP30]] monotonic monotonic, align 2
-// CHECK3-NEXT: [[TMP32]] = extractvalue { i16, i1 } [[TMP31]], 0
-// CHECK3-NEXT: [[TMP33:%.*]] = extractvalue { i16, i1 } [[TMP31]], 1
-// CHECK3-NEXT: br i1 [[TMP33]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK3-NEXT: [[TMP29:%.*]] = load i16, ptr [[ATOMIC_TEMP]], align 2
+// CHECK3-NEXT: [[TMP30:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i16 [[TMP26]], i16 [[TMP29]] monotonic monotonic, align 2
+// CHECK3-NEXT: [[TMP31]] = extractvalue { i16, i1 } [[TMP30]], 0
+// CHECK3-NEXT: [[TMP32:%.*]] = extractvalue { i16, i1 } [[TMP30]], 1
+// CHECK3-NEXT: br i1 [[TMP32]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK3: atomic_exit:
// CHECK3-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT18]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i32 1
// CHECK3-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT19]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], i32 1
-// CHECK3-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT18]], [[TMP25]]
+// CHECK3-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT18]], [[TMP24]]
// CHECK3-NEXT: br i1 [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_BODY9]]
// CHECK3: omp.arraycpy.done21:
// CHECK3-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK3: .omp.reduction.default:
-// CHECK3-NEXT: [[TMP34:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK3-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP34]])
+// CHECK3-NEXT: [[TMP33:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK3-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP33]])
// CHECK3-NEXT: ret void
//
//
@@ -1815,15 +1813,15 @@ int main() {
// CHECK3-NEXT: [[BF_SET:%.*]] = or i8 [[BF_CLEAR]], 0
// CHECK3-NEXT: store i8 [[BF_SET]], ptr [[B]], align 4
// CHECK3-NEXT: [[C:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 2
-// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META5:![0-9]+]], !align [[META6:![0-9]+]]
+// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META4:![0-9]+]], !align [[META5:![0-9]+]]
// CHECK3-NEXT: store ptr [[TMP0]], ptr [[C]], align 8
// CHECK3-NEXT: [[A3:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 0
// CHECK3-NEXT: store ptr [[A3]], ptr [[A2]], align 8
// CHECK3-NEXT: [[C6:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 2
-// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[C6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[C6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[C5]], align 8
-// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 4, ptr @_ZN2SSC2ERi.omp_outlined, ptr [[THIS1]], ptr [[TMP2]], ptr [[B4]], ptr [[TMP3]])
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[B4]], align 4
// CHECK3-NEXT: [[B7:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 1
@@ -1861,27 +1859,27 @@ int main() {
// CHECK3-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK3-NEXT: store ptr [[C]], ptr [[C_ADDR]], align 8
// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
-// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
// CHECK3-NEXT: store ptr [[TMP3]], ptr [[_TMP1]], align 8
-// CHECK3-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store i32 0, ptr [[A2]], align 4
// CHECK3-NEXT: store ptr [[A2]], ptr [[_TMP3]], align 8
// CHECK3-NEXT: store i32 0, ptr [[B4]], align 4
-// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store i32 0, ptr [[C5]], align 4
// CHECK3-NEXT: store ptr [[C5]], ptr [[_TMP6]], align 8
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[REF_TMP]], i32 0, i32 0
// CHECK3-NEXT: store ptr [[TMP0]], ptr [[TMP6]], align 8
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[REF_TMP]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP8:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP8:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store ptr [[TMP8]], ptr [[TMP7]], align 8
// CHECK3-NEXT: [[TMP9:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[REF_TMP]], i32 0, i32 2
// CHECK3-NEXT: store ptr [[B4]], ptr [[TMP9]], align 8
// CHECK3-NEXT: [[TMP10:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[REF_TMP]], i32 0, i32 3
-// CHECK3-NEXT: [[TMP11:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP11:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store ptr [[TMP11]], ptr [[TMP10]], align 8
// CHECK3-NEXT: call void @_ZZN2SSC1ERiENKUlvE_clEv(ptr noundef nonnull align 8 dereferenceable(32) [[REF_TMP]])
// CHECK3-NEXT: [[TMP12:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
@@ -1933,26 +1931,26 @@ int main() {
// CHECK3-NEXT: [[TMP0:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0:%.*]], ptr [[THIS1]], i32 0, i32 0
// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[TMP0]], align 8
// CHECK3-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[THIS1]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[TMP2]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[TMP2]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[TMP3]], align 4
// CHECK3-NEXT: [[INC:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK3-NEXT: store i32 [[INC]], ptr [[TMP3]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[THIS1]], i32 0, i32 2
-// CHECK3-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP6:%.*]] = load ptr, ptr [[TMP5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[DEC:%.*]] = add nsw i32 [[TMP7]], -1
// CHECK3-NEXT: store i32 [[DEC]], ptr [[TMP6]], align 4
// CHECK3-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[THIS1]], i32 0, i32 3
-// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[TMP8]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[TMP8]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
// CHECK3-NEXT: [[DIV:%.*]] = sdiv i32 [[TMP10]], 1
// CHECK3-NEXT: store i32 [[DIV]], ptr [[TMP9]], align 4
// CHECK3-NEXT: [[TMP11:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[THIS1]], i32 0, i32 1
-// CHECK3-NEXT: [[TMP12:%.*]] = load ptr, ptr [[TMP11]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP12:%.*]] = load ptr, ptr [[TMP11]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[THIS1]], i32 0, i32 2
-// CHECK3-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP15:%.*]] = getelementptr inbounds nuw [[CLASS_ANON_0]], ptr [[THIS1]], i32 0, i32 3
-// CHECK3-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 4, ptr @_ZZN2SSC1ERiENKUlvE_clEv.omp_outlined, ptr [[TMP1]], ptr [[TMP12]], ptr [[TMP14]], ptr [[TMP16]])
// CHECK3-NEXT: ret void
//
@@ -2017,26 +2015,26 @@ int main() {
// CHECK3-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK3-NEXT: store ptr [[C]], ptr [[C_ADDR]], align 8
// CHECK3-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
-// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK3-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK3-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
// CHECK3-NEXT: store ptr [[TMP3]], ptr [[_TMP1]], align 8
-// CHECK3-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store i32 -1, ptr [[A2]], align 4
// CHECK3-NEXT: store ptr [[A2]], ptr [[_TMP3]], align 8
// CHECK3-NEXT: store i32 -1, ptr [[B4]], align 4
-// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: store i32 -1, ptr [[C5]], align 4
// CHECK3-NEXT: store ptr [[C5]], ptr [[_TMP6]], align 8
-// CHECK3-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[TMP6]], align 4
// CHECK3-NEXT: [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK3-NEXT: store i32 [[INC]], ptr [[TMP6]], align 4
// CHECK3-NEXT: [[TMP8:%.*]] = load i32, ptr [[B4]], align 4
// CHECK3-NEXT: [[DEC:%.*]] = add nsw i32 [[TMP8]], -1
// CHECK3-NEXT: store i32 [[DEC]], ptr [[B4]], align 4
-// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK3-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
// CHECK3-NEXT: [[DIV:%.*]] = sdiv i32 [[TMP10]], 1
// CHECK3-NEXT: store i32 [[DIV]], ptr [[TMP9]], align 4
@@ -2148,99 +2146,98 @@ int main() {
// CHECK4-NEXT: [[TMP2:%.*]] = ptrtoaddr ptr [[ARRAYIDX1]] to i64
// CHECK4-NEXT: [[TMP3:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
// CHECK4-NEXT: [[TMP4:%.*]] = sub i64 [[TMP2]], [[TMP3]]
-// CHECK4-NEXT: [[TMP5:%.*]] = sdiv exact i64 [[TMP4]], 2
-// CHECK4-NEXT: [[TMP6:%.*]] = add nuw i64 [[TMP5]], 1
-// CHECK4-NEXT: [[TMP7:%.*]] = mul nuw i64 [[TMP6]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK4-NEXT: [[TMP8:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK4-NEXT: store ptr [[TMP8]], ptr [[SAVED_STACK]], align 8
+// CHECK4-NEXT: [[TMP5:%.*]] = add nuw i64 [[TMP4]], 2
+// CHECK4-NEXT: [[TMP6:%.*]] = udiv exact i64 [[TMP5]], 2
+// CHECK4-NEXT: [[TMP7:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK4-NEXT: store ptr [[TMP7]], ptr [[SAVED_STACK]], align 8
// CHECK4-NEXT: [[VLA:%.*]] = alloca i16, i64 [[TMP6]], align 16
// CHECK4-NEXT: store i64 [[TMP6]], ptr [[__VLA_EXPR0]], align 8
-// CHECK4-NEXT: [[TMP9:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP6]]
-// CHECK4-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP9]]
+// CHECK4-NEXT: [[TMP8:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP6]]
+// CHECK4-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP8]]
// CHECK4-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK4: omp.arrayinit.body:
// CHECK4-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK4-NEXT: store i16 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 2
// CHECK4-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK4-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP9]]
+// CHECK4-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP8]]
// CHECK4-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK4: omp.arrayinit.done:
-// CHECK4-NEXT: [[TMP10:%.*]] = load ptr, ptr [[X_ADDR]], align 8
-// CHECK4-NEXT: [[TMP11:%.*]] = ptrtoaddr ptr [[TMP10]] to i64
-// CHECK4-NEXT: [[TMP12:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
-// CHECK4-NEXT: [[TMP13:%.*]] = sub i64 [[TMP11]], [[TMP12]]
-// CHECK4-NEXT: [[TMP14:%.*]] = sdiv exact i64 [[TMP13]], 2
-// CHECK4-NEXT: [[TMP15:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP14]]
-// CHECK4-NEXT: store ptr [[TMP15]], ptr [[TMP]], align 8
-// CHECK4-NEXT: [[TMP16:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK4-NEXT: store ptr [[VLA]], ptr [[TMP16]], align 8
-// CHECK4-NEXT: [[TMP17:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK4-NEXT: [[TMP18:%.*]] = inttoptr i64 [[TMP6]] to ptr
-// CHECK4-NEXT: store ptr [[TMP18]], ptr [[TMP17]], align 8
-// CHECK4-NEXT: [[TMP19:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK4-NEXT: [[TMP20:%.*]] = load i32, ptr [[TMP19]], align 4
-// CHECK4-NEXT: [[TMP21:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1:[0-9]+]], i32 [[TMP20]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_Z14foo_array_sectPs.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK4-NEXT: switch i32 [[TMP21]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK4-NEXT: [[TMP9:%.*]] = load ptr, ptr [[X_ADDR]], align 8
+// CHECK4-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[TMP9]] to i64
+// CHECK4-NEXT: [[TMP11:%.*]] = ptrtoaddr ptr [[ARRAYIDX]] to i64
+// CHECK4-NEXT: [[TMP12:%.*]] = sub i64 [[TMP10]], [[TMP11]]
+// CHECK4-NEXT: [[TMP13:%.*]] = sdiv exact i64 [[TMP12]], 2
+// CHECK4-NEXT: [[TMP14:%.*]] = getelementptr i16, ptr [[VLA]], i64 [[TMP13]]
+// CHECK4-NEXT: store ptr [[TMP14]], ptr [[TMP]], align 8
+// CHECK4-NEXT: [[TMP15:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK4-NEXT: store ptr [[VLA]], ptr [[TMP15]], align 8
+// CHECK4-NEXT: [[TMP16:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK4-NEXT: [[TMP17:%.*]] = inttoptr i64 [[TMP6]] to ptr
+// CHECK4-NEXT: store ptr [[TMP17]], ptr [[TMP16]], align 8
+// CHECK4-NEXT: [[TMP18:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK4-NEXT: [[TMP19:%.*]] = load i32, ptr [[TMP18]], align 4
+// CHECK4-NEXT: [[TMP20:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1:[0-9]+]], i32 [[TMP19]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @_Z14foo_array_sectPs.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK4-NEXT: switch i32 [[TMP20]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
// CHECK4-NEXT: i32 1, label [[DOTOMP_REDUCTION_CASE1:%.*]]
// CHECK4-NEXT: i32 2, label [[DOTOMP_REDUCTION_CASE2:%.*]]
// CHECK4-NEXT: ]
// CHECK4: .omp.reduction.case1:
-// CHECK4-NEXT: [[TMP22:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
-// CHECK4-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP22]]
+// CHECK4-NEXT: [[TMP21:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
+// CHECK4-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP21]]
// CHECK4-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE7:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK4: omp.arraycpy.body:
// CHECK4-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK4-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST2:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT5:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK4-NEXT: [[TMP23:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
-// CHECK4-NEXT: [[CONV:%.*]] = sext i16 [[TMP23]] to i32
-// CHECK4-NEXT: [[TMP24:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
-// CHECK4-NEXT: [[CONV3:%.*]] = sext i16 [[TMP24]] to i32
+// CHECK4-NEXT: [[TMP22:%.*]] = load i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
+// CHECK4-NEXT: [[CONV:%.*]] = sext i16 [[TMP22]] to i32
+// CHECK4-NEXT: [[TMP23:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 2
+// CHECK4-NEXT: [[CONV3:%.*]] = sext i16 [[TMP23]] to i32
// CHECK4-NEXT: [[ADD:%.*]] = add nsw i32 [[CONV]], [[CONV3]]
// CHECK4-NEXT: [[CONV4:%.*]] = trunc i32 [[ADD]] to i16
// CHECK4-NEXT: store i16 [[CONV4]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], align 2
// CHECK4-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT5]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST2]], i32 1
// CHECK4-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK4-NEXT: [[OMP_ARRAYCPY_DONE6:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP22]]
+// CHECK4-NEXT: [[OMP_ARRAYCPY_DONE6:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT5]], [[TMP21]]
// CHECK4-NEXT: br i1 [[OMP_ARRAYCPY_DONE6]], label [[OMP_ARRAYCPY_DONE7]], label [[OMP_ARRAYCPY_BODY]]
// CHECK4: omp.arraycpy.done7:
-// CHECK4-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP20]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK4-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP19]], ptr @.gomp_critical_user_.reduction.var)
// CHECK4-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK4: .omp.reduction.case2:
-// CHECK4-NEXT: [[TMP25:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
-// CHECK4-NEXT: [[OMP_ARRAYCPY_ISEMPTY8:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP25]]
+// CHECK4-NEXT: [[TMP24:%.*]] = getelementptr i16, ptr [[ARRAYIDX]], i64 [[TMP6]]
+// CHECK4-NEXT: [[OMP_ARRAYCPY_ISEMPTY8:%.*]] = icmp eq ptr [[ARRAYIDX]], [[TMP24]]
// CHECK4-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY8]], label [[OMP_ARRAYCPY_DONE21:%.*]], label [[OMP_ARRAYCPY_BODY9:%.*]]
// CHECK4: omp.arraycpy.body9:
// CHECK4-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST10:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT19:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK4-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST11:%.*]] = phi ptr [ [[ARRAYIDX]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT18:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK4-NEXT: [[TMP26:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
-// CHECK4-NEXT: [[CONV12:%.*]] = sext i16 [[TMP26]] to i32
+// CHECK4-NEXT: [[TMP25:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
+// CHECK4-NEXT: [[CONV12:%.*]] = sext i16 [[TMP25]] to i32
// CHECK4-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]] monotonic, align 2
// CHECK4-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK4: atomic_cont:
-// CHECK4-NEXT: [[TMP27:%.*]] = phi i16 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY9]] ], [ [[TMP32:%.*]], [[ATOMIC_CONT]] ]
-// CHECK4-NEXT: store i16 [[TMP27]], ptr [[_TMP13]], align 2
-// CHECK4-NEXT: [[TMP28:%.*]] = load i16, ptr [[_TMP13]], align 2
-// CHECK4-NEXT: [[CONV14:%.*]] = sext i16 [[TMP28]] to i32
-// CHECK4-NEXT: [[TMP29:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
-// CHECK4-NEXT: [[CONV15:%.*]] = sext i16 [[TMP29]] to i32
+// CHECK4-NEXT: [[TMP26:%.*]] = phi i16 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY9]] ], [ [[TMP31:%.*]], [[ATOMIC_CONT]] ]
+// CHECK4-NEXT: store i16 [[TMP26]], ptr [[_TMP13]], align 2
+// CHECK4-NEXT: [[TMP27:%.*]] = load i16, ptr [[_TMP13]], align 2
+// CHECK4-NEXT: [[CONV14:%.*]] = sext i16 [[TMP27]] to i32
+// CHECK4-NEXT: [[TMP28:%.*]] = load i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], align 2
+// CHECK4-NEXT: [[CONV15:%.*]] = sext i16 [[TMP28]] to i32
// CHECK4-NEXT: [[ADD16:%.*]] = add nsw i32 [[CONV14]], [[CONV15]]
// CHECK4-NEXT: [[CONV17:%.*]] = trunc i32 [[ADD16]] to i16
// CHECK4-NEXT: store i16 [[CONV17]], ptr [[ATOMIC_TEMP]], align 2
-// CHECK4-NEXT: [[TMP30:%.*]] = load i16, ptr [[ATOMIC_TEMP]], align 2
-// CHECK4-NEXT: [[TMP31:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i16 [[TMP27]], i16 [[TMP30]] monotonic monotonic, align 2
-// CHECK4-NEXT: [[TMP32]] = extractvalue { i16, i1 } [[TMP31]], 0
-// CHECK4-NEXT: [[TMP33:%.*]] = extractvalue { i16, i1 } [[TMP31]], 1
-// CHECK4-NEXT: br i1 [[TMP33]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK4-NEXT: [[TMP29:%.*]] = load i16, ptr [[ATOMIC_TEMP]], align 2
+// CHECK4-NEXT: [[TMP30:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i16 [[TMP26]], i16 [[TMP29]] monotonic monotonic, align 2
+// CHECK4-NEXT: [[TMP31]] = extractvalue { i16, i1 } [[TMP30]], 0
+// CHECK4-NEXT: [[TMP32:%.*]] = extractvalue { i16, i1 } [[TMP30]], 1
+// CHECK4-NEXT: br i1 [[TMP32]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK4: atomic_exit:
// CHECK4-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT18]] = getelementptr i16, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST11]], i32 1
// CHECK4-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT19]] = getelementptr i16, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST10]], i32 1
-// CHECK4-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT18]], [[TMP25]]
+// CHECK4-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT18]], [[TMP24]]
// CHECK4-NEXT: br i1 [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_BODY9]]
// CHECK4: omp.arraycpy.done21:
// CHECK4-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK4: .omp.reduction.default:
-// CHECK4-NEXT: [[TMP34:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK4-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP34]])
+// CHECK4-NEXT: [[TMP33:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK4-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP33]])
// CHECK4-NEXT: ret void
//
//
@@ -2307,7 +2304,7 @@ int main() {
//
//
// CHECK4-LABEL: define {{[^@]+}}@__main_block_invoke
-// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR8:[0-9]+]] {
+// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca ptr, align 8
// CHECK4-NEXT: [[BLOCK_ADDR:%.*]] = alloca ptr, align 8
@@ -2329,7 +2326,7 @@ int main() {
// CHECK4-NEXT: store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK4-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK4-NEXT: store ptr [[G]], ptr [[G_ADDR]], align 8
-// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[G_ADDR]], align 8, !nonnull [[META5:![0-9]+]], !align [[META6:![0-9]+]]
+// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[G_ADDR]], align 8, !nonnull [[META4:![0-9]+]], !align [[META5:![0-9]+]]
// CHECK4-NEXT: store i32 0, ptr [[G1]], align 128
// CHECK4-NEXT: store i32 1, ptr [[G1]], align 128
// CHECK4-NEXT: [[BLOCK_ISA:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, [96 x i8], i32 }>, ptr [[BLOCK]], i32 0, i32 0
@@ -2373,7 +2370,7 @@ int main() {
//
//
// CHECK4-LABEL: define {{[^@]+}}@g_block_invoke
-// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR8]] {
+// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca ptr, align 8
// CHECK4-NEXT: [[BLOCK_ADDR:%.*]] = alloca ptr, align 8
@@ -2423,15 +2420,15 @@ int main() {
// CHECK4-NEXT: [[BF_SET:%.*]] = or i8 [[BF_CLEAR]], 0
// CHECK4-NEXT: store i8 [[BF_SET]], ptr [[B]], align 4
// CHECK4-NEXT: [[C:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 2
-// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP0]], ptr [[C]], align 8
// CHECK4-NEXT: [[A3:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 0
// CHECK4-NEXT: store ptr [[A3]], ptr [[A2]], align 8
// CHECK4-NEXT: [[C6:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 2
-// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr [[C6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr [[C6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP1]], ptr [[C5]], align 8
-// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[A2]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 4, ptr @_ZN2SSC2ERi.omp_outlined, ptr [[THIS1]], ptr [[TMP2]], ptr [[B4]], ptr [[TMP3]])
// CHECK4-NEXT: [[TMP4:%.*]] = load i32, ptr [[B4]], align 4
// CHECK4-NEXT: [[B7:%.*]] = getelementptr inbounds nuw [[STRUCT_SS]], ptr [[THIS1]], i32 0, i32 1
@@ -2469,16 +2466,16 @@ int main() {
// CHECK4-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK4-NEXT: store ptr [[C]], ptr [[C_ADDR]], align 8
// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
-// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
// CHECK4-NEXT: store ptr [[TMP3]], ptr [[_TMP1]], align 8
-// CHECK4-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store i32 0, ptr [[A2]], align 4
// CHECK4-NEXT: store ptr [[A2]], ptr [[_TMP3]], align 8
// CHECK4-NEXT: store i32 0, ptr [[B4]], align 4
-// CHECK4-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store i32 0, ptr [[C5]], align 4
// CHECK4-NEXT: store ptr [[C5]], ptr [[_TMP6]], align 8
// CHECK4-NEXT: [[BLOCK_ISA:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[BLOCK]], i32 0, i32 0
@@ -2494,13 +2491,13 @@ int main() {
// CHECK4-NEXT: [[BLOCK_CAPTURED_THIS_ADDR:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[BLOCK]], i32 0, i32 5
// CHECK4-NEXT: store ptr [[TMP0]], ptr [[BLOCK_CAPTURED_THIS_ADDR]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURED:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[BLOCK]], i32 0, i32 6
-// CHECK4-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP6]], ptr [[BLOCK_CAPTURED]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURED7:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[BLOCK]], i32 0, i32 8
// CHECK4-NEXT: [[TMP7:%.*]] = load i32, ptr [[B4]], align 4
// CHECK4-NEXT: store i32 [[TMP7]], ptr [[BLOCK_CAPTURED7]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURED8:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[BLOCK]], i32 0, i32 7
-// CHECK4-NEXT: [[TMP8:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP8:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP8]], ptr [[BLOCK_CAPTURED8]], align 8
// CHECK4-NEXT: [[TMP9:%.*]] = getelementptr inbounds nuw [[STRUCT___BLOCK_LITERAL_GENERIC:%.*]], ptr [[BLOCK]], i32 0, i32 3
// CHECK4-NEXT: [[TMP10:%.*]] = load ptr, ptr [[TMP9]], align 8
@@ -2546,7 +2543,7 @@ int main() {
//
//
// CHECK4-LABEL: define {{[^@]+}}@g_block_invoke_2
-// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR8]] {
+// CHECK4-SAME: (ptr noundef [[DOTBLOCK_DESCRIPTOR:%.*]]) #[[ATTR2]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[DOTBLOCK_DESCRIPTOR_ADDR:%.*]] = alloca ptr, align 8
// CHECK4-NEXT: [[BLOCK_ADDR:%.*]] = alloca ptr, align 8
@@ -2555,7 +2552,7 @@ int main() {
// CHECK4-NEXT: [[BLOCK_CAPTURED_THIS:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 5
// CHECK4-NEXT: [[THIS:%.*]] = load ptr, ptr [[BLOCK_CAPTURED_THIS]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 6
-// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: [[TMP1:%.*]] = load i32, ptr [[TMP0]], align 4
// CHECK4-NEXT: [[INC:%.*]] = add nsw i32 [[TMP1]], 1
// CHECK4-NEXT: store i32 [[INC]], ptr [[TMP0]], align 4
@@ -2564,15 +2561,15 @@ int main() {
// CHECK4-NEXT: [[DEC:%.*]] = add nsw i32 [[TMP2]], -1
// CHECK4-NEXT: store i32 [[DEC]], ptr [[BLOCK_CAPTURE_ADDR1]], align 8
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR2:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 7
-// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR2]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR2]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: [[TMP4:%.*]] = load i32, ptr [[TMP3]], align 4
// CHECK4-NEXT: [[DIV:%.*]] = sdiv i32 [[TMP4]], 1
// CHECK4-NEXT: store i32 [[DIV]], ptr [[TMP3]], align 4
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR3:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 6
-// CHECK4-NEXT: [[TMP5:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP5:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR4:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 8
// CHECK4-NEXT: [[BLOCK_CAPTURE_ADDR5:%.*]] = getelementptr inbounds nuw <{ ptr, i32, i32, ptr, ptr, ptr, ptr, ptr, i32 }>, ptr [[DOTBLOCK_DESCRIPTOR]], i32 0, i32 7
-// CHECK4-NEXT: [[TMP6:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR5]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP6:%.*]] = load ptr, ptr [[BLOCK_CAPTURE_ADDR5]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB2]], i32 4, ptr @g_block_invoke_2.omp_outlined, ptr [[THIS]], ptr [[TMP5]], ptr [[BLOCK_CAPTURE_ADDR4]], ptr [[TMP6]])
// CHECK4-NEXT: ret void
//
@@ -2601,26 +2598,26 @@ int main() {
// CHECK4-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
// CHECK4-NEXT: store ptr [[C]], ptr [[C_ADDR]], align 8
// CHECK4-NEXT: [[TMP0:%.*]] = load ptr, ptr [[THIS_ADDR]], align 8
-// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
-// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP1:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK4-NEXT: [[TMP2:%.*]] = load ptr, ptr [[B_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
+// CHECK4-NEXT: [[TMP3:%.*]] = load ptr, ptr [[C_ADDR]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store ptr [[TMP1]], ptr [[TMP]], align 8
// CHECK4-NEXT: store ptr [[TMP3]], ptr [[_TMP1]], align 8
-// CHECK4-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP4:%.*]] = load ptr, ptr [[TMP]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store i32 0, ptr [[A2]], align 4
// CHECK4-NEXT: store ptr [[A2]], ptr [[_TMP3]], align 8
// CHECK4-NEXT: store i32 0, ptr [[B4]], align 4
-// CHECK4-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP5:%.*]] = load ptr, ptr [[_TMP1]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: store i32 0, ptr [[C5]], align 4
// CHECK4-NEXT: store ptr [[C5]], ptr [[_TMP6]], align 8
-// CHECK4-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP6:%.*]] = load ptr, ptr [[_TMP3]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: [[TMP7:%.*]] = load i32, ptr [[TMP6]], align 4
// CHECK4-NEXT: [[INC:%.*]] = add nsw i32 [[TMP7]], 1
// CHECK4-NEXT: store i32 [[INC]], ptr [[TMP6]], align 4
// CHECK4-NEXT: [[TMP8:%.*]] = load i32, ptr [[B4]], align 4
// CHECK4-NEXT: [[DEC:%.*]] = add nsw i32 [[TMP8]], -1
// CHECK4-NEXT: store i32 [[DEC]], ptr [[B4]], align 4
-// CHECK4-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META5]], !align [[META6]]
+// CHECK4-NEXT: [[TMP9:%.*]] = load ptr, ptr [[_TMP6]], align 8, !nonnull [[META4]], !align [[META5]]
// CHECK4-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
// CHECK4-NEXT: [[DIV:%.*]] = sdiv i32 [[TMP10]], 1
// CHECK4-NEXT: store i32 [[DIV]], ptr [[TMP9]], align 4
diff --git a/clang/test/OpenMP/parallel_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_reduction_task_codegen.cpp
index 334319c27b143..46415d4557eb0 100644
--- a/clang/test/OpenMP/parallel_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_reduction_task_codegen.cpp
@@ -69,7 +69,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[ARGC1]], align 4
// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
// CHECK1-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP1]], i64 0
@@ -86,180 +86,178 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 0, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP48]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP49]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP51:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP51]], ptr [[TMP50]], align 8
-// CHECK1-NEXT: [[TMP52:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP53:%.*]] = load i32, ptr [[TMP52]], align 4
-// CHECK1-NEXT: [[TMP54:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP53]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP54]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP55]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP57:%.*]] = load ptr, ptr [[TMP56]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP57]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP54]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP58]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP60]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load i32, ptr [[TMP61]], align 4
-// CHECK1-NEXT: [[TMP63:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP62]], ptr [[TMP54]])
-// CHECK1-NEXT: [[TMP64:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP65:%.*]] = load i32, ptr [[TMP64]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP65]], i32 0)
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP66]], align 8
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP67]], align 8
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP69:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP69]], ptr [[TMP68]], align 8
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[TMP70]], align 4
-// CHECK1-NEXT: [[TMP72:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP71]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP72]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 0, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP46]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP47]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP49:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: store ptr [[TMP49]], ptr [[TMP48]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[TMP50]], align 4
+// CHECK1-NEXT: [[TMP52:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP51]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP52]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP53]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP55:%.*]] = load ptr, ptr [[TMP54]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP55]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP52]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP56]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP58:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP58]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load i32, ptr [[TMP59]], align 4
+// CHECK1-NEXT: [[TMP61:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP60]], ptr [[TMP52]])
+// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP63]], i32 0)
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP64]], align 8
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP65]], align 8
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP67:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP67]], ptr [[TMP66]], align 8
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
+// CHECK1-NEXT: [[TMP70:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP69]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP70]], 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: [[TMP73:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP73]], [[TMP74]]
+// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP71]], [[TMP72]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP75:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP75]]
+// CHECK1-NEXT: [[TMP73:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP73]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE18:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST12:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT16:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP76:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP76]] to i32
-// CHECK1-NEXT: [[TMP77:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP77]] to i32
+// CHECK1-NEXT: [[TMP74:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP74]] to i32
+// CHECK1-NEXT: [[TMP75:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP75]] to i32
// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[CONV]], [[CONV13]]
// CHECK1-NEXT: [[CONV15:%.*]] = trunc i32 [[ADD14]] to i8
// CHECK1-NEXT: store i8 [[CONV15]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT16]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP75]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP73]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE17]], label [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done18:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP71]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP69]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP79:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP78]] monotonic, align 4
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP80]]
+// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP77:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP76]] monotonic, align 4
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP78]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY19]], label [[OMP_ARRAYCPY_DONE32:%.*]], label [[OMP_ARRAYCPY_BODY20:%.*]]
// CHECK1: omp.arraycpy.body20:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST21:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT30:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST22:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT29:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP81]] to i32
+// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP79]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP82:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP87:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP82]], ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP83]] to i32
-// CHECK1-NEXT: [[TMP84:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP84]] to i32
+// CHECK1-NEXT: [[TMP80:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP85:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP80]], ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP81]] to i32
+// CHECK1-NEXT: [[TMP82:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP82]] to i32
// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[CONV25]], [[CONV26]]
// CHECK1-NEXT: [[CONV28:%.*]] = trunc i32 [[ADD27]] to i8
// CHECK1-NEXT: store i8 [[CONV28]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP85:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP86:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP82]], i8 [[TMP85]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP87]] = extractvalue { i8, i1 } [[TMP86]], 0
-// CHECK1-NEXT: [[TMP88:%.*]] = extractvalue { i8, i1 } [[TMP86]], 1
-// CHECK1-NEXT: br i1 [[TMP88]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP84:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP80]], i8 [[TMP83]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP85]] = extractvalue { i8, i1 } [[TMP84]], 0
+// CHECK1-NEXT: [[TMP86:%.*]] = extractvalue { i8, i1 } [[TMP84]], 1
+// CHECK1-NEXT: br i1 [[TMP86]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT29]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT30]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP80]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP78]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE31]], label [[OMP_ARRAYCPY_DONE32]], label [[OMP_ARRAYCPY_BODY20]]
// CHECK1: omp.arraycpy.done32:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP89:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP89]])
+// CHECK1-NEXT: [[TMP87:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP87]])
// CHECK1-NEXT: ret void
//
//
@@ -383,31 +381,31 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]]
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META4:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]], !inline_history [[META14:![0-9]+]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13]]
// CHECK1-NEXT: [[TMP17:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP15]], ptr [[TMP14]])
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[TMP18]], align 8
// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[TMP22]], align 4
// CHECK1-NEXT: [[TMP24:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[LB_ADD_LEN_I:%.*]] = add nsw i64 -1, [[TMP24]]
@@ -420,19 +418,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
-// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
+// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
index 0c92eae259c65..965c3037c2d11 100644
--- a/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
@@ -74,7 +74,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_SECTIONS_ST_]], align 4
@@ -95,213 +95,211 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2:[0-9]+]], i32 [[TMP49]], i32 34, ptr [[DOTOMP_SECTIONS_IL_]], ptr [[DOTOMP_SECTIONS_LB_]], ptr [[DOTOMP_SECTIONS_UB_]], ptr [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP50:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
-// CHECK1-NEXT: [[TMP51:%.*]] = icmp slt i32 [[TMP50]], 0
-// CHECK1-NEXT: [[TMP52:%.*]] = select i1 [[TMP51]], i32 [[TMP50]], i32 0
-// CHECK1-NEXT: store i32 [[TMP52]], ptr [[DOTOMP_SECTIONS_UB_]], align 4
-// CHECK1-NEXT: [[TMP53:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_LB_]], align 4
-// CHECK1-NEXT: store i32 [[TMP53]], ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2:[0-9]+]], i32 [[TMP47]], i32 34, ptr [[DOTOMP_SECTIONS_IL_]], ptr [[DOTOMP_SECTIONS_LB_]], ptr [[DOTOMP_SECTIONS_UB_]], ptr [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP48:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
+// CHECK1-NEXT: [[TMP49:%.*]] = icmp slt i32 [[TMP48]], 0
+// CHECK1-NEXT: [[TMP50:%.*]] = select i1 [[TMP49]], i32 [[TMP48]], i32 0
+// CHECK1-NEXT: store i32 [[TMP50]], ptr [[DOTOMP_SECTIONS_UB_]], align 4
+// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_LB_]], align 4
+// CHECK1-NEXT: store i32 [[TMP51]], ptr [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP54:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
-// CHECK1-NEXT: [[TMP55:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sle i32 [[TMP54]], [[TMP55]]
+// CHECK1-NEXT: [[TMP52:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: [[TMP53:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sle i32 [[TMP52]], [[TMP53]]
// CHECK1-NEXT: br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP56:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
-// CHECK1-NEXT: switch i32 [[TMP56]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
+// CHECK1-NEXT: [[TMP54:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: switch i32 [[TMP54]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK1-NEXT: i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK1-NEXT: ]
// CHECK1: .omp.sections.case:
-// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP57]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP60]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load i32, ptr [[TMP61]], align 4
-// CHECK1-NEXT: [[TMP63:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP62]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP63]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP64]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[TMP65]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP66]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP63]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP67]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP69]], ptr [[TMP68]], align 8
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[TMP70]], align 4
-// CHECK1-NEXT: [[TMP72:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP71]], ptr [[TMP63]])
+// CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP55]], align 8
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP58:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: store ptr [[TMP58]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load i32, ptr [[TMP59]], align 4
+// CHECK1-NEXT: [[TMP61:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP60]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP61]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP62]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP64:%.*]] = load ptr, ptr [[TMP63]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP64]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP61]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP65]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP67]], ptr [[TMP66]], align 8
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
+// CHECK1-NEXT: [[TMP70:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP69]], ptr [[TMP61]])
// CHECK1-NEXT: br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK1: .omp.sections.exit:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1: omp.inner.for.inc:
-// CHECK1-NEXT: [[TMP73:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
-// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP73]], 1
+// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP71]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK1: omp.inner.for.end:
+// CHECK1-NEXT: [[TMP72:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP73:%.*]] = load i32, ptr [[TMP72]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP73]])
// CHECK1-NEXT: [[TMP74:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP75:%.*]] = load i32, ptr [[TMP74]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP75]])
-// CHECK1-NEXT: [[TMP76:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP77:%.*]] = load i32, ptr [[TMP76]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP77]], i32 1)
-// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP78]], align 8
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP81:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP81]], ptr [[TMP80]], align 8
-// CHECK1-NEXT: [[TMP82:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP83:%.*]] = load i32, ptr [[TMP82]], align 4
-// CHECK1-NEXT: [[TMP84:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3:[0-9]+]], i32 [[TMP83]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP84]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP75]], i32 1)
+// CHECK1-NEXT: [[TMP76:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP76]], align 8
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP79:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP79]], ptr [[TMP78]], align 8
+// CHECK1-NEXT: [[TMP80:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP81:%.*]] = load i32, ptr [[TMP80]], align 4
+// CHECK1-NEXT: [[TMP82:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3:[0-9]+]], i32 [[TMP81]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP82]], 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: [[TMP85:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP86:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP85]], [[TMP86]]
+// CHECK1-NEXT: [[TMP83:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP83]], [[TMP84]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP87:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP87]]
+// CHECK1-NEXT: [[TMP85:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP85]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE18:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST12:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT16:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP88]] to i32
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP89]] to i32
+// CHECK1-NEXT: [[TMP86:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP86]] to i32
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP87]] to i32
// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[CONV]], [[CONV13]]
// CHECK1-NEXT: [[CONV15:%.*]] = trunc i32 [[ADD14]] to i8
// CHECK1-NEXT: store i8 [[CONV15]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT16]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP87]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP85]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE17]], label [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done18:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP83]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP81]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP90:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP91:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP90]] monotonic, align 4
-// CHECK1-NEXT: [[TMP92:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP92]]
+// CHECK1-NEXT: [[TMP88:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP89:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP88]] monotonic, align 4
+// CHECK1-NEXT: [[TMP90:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP90]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY19]], label [[OMP_ARRAYCPY_DONE32:%.*]], label [[OMP_ARRAYCPY_BODY20:%.*]]
// CHECK1: omp.arraycpy.body20:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST21:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT30:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST22:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT29:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP93:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP93]] to i32
+// CHECK1-NEXT: [[TMP91:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP91]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP94:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP99:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP94]], ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP95]] to i32
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP96]] to i32
+// CHECK1-NEXT: [[TMP92:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP97:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP92]], ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[TMP93:%.*]] = load i8, ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP93]] to i32
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP94]] to i32
// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[CONV25]], [[CONV26]]
// CHECK1-NEXT: [[CONV28:%.*]] = trunc i32 [[ADD27]] to i8
// CHECK1-NEXT: store i8 [[CONV28]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP97:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP98:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP94]], i8 [[TMP97]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP99]] = extractvalue { i8, i1 } [[TMP98]], 0
-// CHECK1-NEXT: [[TMP100:%.*]] = extractvalue { i8, i1 } [[TMP98]], 1
-// CHECK1-NEXT: br i1 [[TMP100]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP96:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP92]], i8 [[TMP95]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP97]] = extractvalue { i8, i1 } [[TMP96]], 0
+// CHECK1-NEXT: [[TMP98:%.*]] = extractvalue { i8, i1 } [[TMP96]], 1
+// CHECK1-NEXT: br i1 [[TMP98]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT29]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT30]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP92]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP90]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE31]], label [[OMP_ARRAYCPY_DONE32]], label [[OMP_ARRAYCPY_BODY20]]
// CHECK1: omp.arraycpy.done32:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP101:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP101]])
+// CHECK1-NEXT: [[TMP99:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP99]])
// CHECK1-NEXT: ret void
//
//
@@ -425,31 +423,31 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]]
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META4:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]], !inline_history [[META14:![0-9]+]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META13]]
// CHECK1-NEXT: [[TMP17:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP15]], ptr [[TMP14]])
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[TMP18]], align 8
// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[TMP22]], align 4
// CHECK1-NEXT: [[TMP24:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[LB_ADD_LEN_I:%.*]] = add nsw i64 -1, [[TMP24]]
@@ -462,19 +460,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META14]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
-// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META14]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
+// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/reduction_implicit_map.cpp b/clang/test/OpenMP/reduction_implicit_map.cpp
index 0c098049b559b..fcdfca7672758 100644
--- a/clang/test/OpenMP/reduction_implicit_map.cpp
+++ b/clang/test/OpenMP/reduction_implicit_map.cpp
@@ -573,30 +573,29 @@ int main()
// CHECK1-NEXT: [[TMP1:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
// CHECK1-NEXT: [[TMP2:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
// CHECK1-NEXT: [[TMP3:%.*]] = sub i64 [[TMP1]], [[TMP2]]
-// CHECK1-NEXT: [[TMP4:%.*]] = sdiv exact i64 [[TMP3]], 8
-// CHECK1-NEXT: [[TMP5:%.*]] = add nuw i64 [[TMP4]], 1
-// CHECK1-NEXT: [[TMP6:%.*]] = mul nuw i64 [[TMP5]], ptrtoint (ptr getelementptr (double, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP7:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP4:%.*]] = add nuw i64 [[TMP3]], 8
+// CHECK1-NEXT: [[TMP5:%.*]] = udiv exact i64 [[TMP4]], 8
+// CHECK1-NEXT: [[TMP6:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP6]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca double, i64 [[TMP5]], align 8
// CHECK1-NEXT: store i64 [[TMP5]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr double, ptr [[VLA]], i64 [[TMP5]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP8]]
+// CHECK1-NEXT: [[TMP7:%.*]] = getelementptr double, ptr [[VLA]], i64 [[TMP5]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP7]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store double 0.000000e+00, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 8
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr double, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP8]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP7]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP9:%.*]] = ptrtoaddr ptr [[TMP0]] to i64
-// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
-// CHECK1-NEXT: [[TMP11:%.*]] = sub i64 [[TMP9]], [[TMP10]]
-// CHECK1-NEXT: [[TMP12:%.*]] = sdiv exact i64 [[TMP11]], 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr double, ptr [[VLA]], i64 [[TMP12]]
+// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[TMP0]] to i64
+// CHECK1-NEXT: [[TMP9:%.*]] = ptrtoaddr ptr [[ARRAYIDX3]] to i64
+// CHECK1-NEXT: [[TMP10:%.*]] = sub i64 [[TMP8]], [[TMP9]]
+// CHECK1-NEXT: [[TMP11:%.*]] = sdiv exact i64 [[TMP10]], 8
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr double, ptr [[VLA]], i64 [[TMP11]]
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [1 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw [10 x [10 x [10 x double]]], ptr [[TMP0]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYDECAY10:%.*]] = getelementptr inbounds [10 x [10 x double]], ptr [[ARRAYIDX9]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw [10 x double], ptr [[ARRAYDECAY10]], i64 2
@@ -607,123 +606,122 @@ int main()
// CHECK1-NEXT: [[ARRAYIDX16:%.*]] = getelementptr inbounds nuw [10 x double], ptr [[ARRAYDECAY15]], i64 5
// CHECK1-NEXT: [[ARRAYDECAY17:%.*]] = getelementptr inbounds [10 x double], ptr [[ARRAYIDX16]], i64 0, i64 0
// CHECK1-NEXT: [[ARRAYIDX18:%.*]] = getelementptr inbounds nuw double, ptr [[ARRAYDECAY17]], i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX13]], ptr [[TMP15]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX18]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX13]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = sdiv exact i64 [[TMP18]], 8
-// CHECK1-NEXT: [[TMP20:%.*]] = add nuw i64 [[TMP19]], 1
-// CHECK1-NEXT: [[TMP21:%.*]] = mul nuw i64 [[TMP20]], ptrtoint (ptr getelementptr (double, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP26]], align 8
-// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[TMP27]], align 4
-// CHECK1-NEXT: [[TMP29:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB2]], i32 [[TMP28]], i32 1, i32 1, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP29]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP31:%.*]] = load i32, ptr [[TMP30]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP31]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP32:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP32]], 9
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX13]], ptr [[TMP14]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[ARRAYIDX18]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX13]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = add nuw i64 [[TMP17]], 8
+// CHECK1-NEXT: [[TMP19:%.*]] = udiv exact i64 [[TMP18]], 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP18]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = load i32, ptr [[TMP25]], align 4
+// CHECK1-NEXT: [[TMP27:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB2]], i32 [[TMP26]], i32 1, i32 1, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP27]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP28]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP29]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP30:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP30]], 9
// 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: [[TMP33:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP31:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP33]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP31]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP34]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP32:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP32]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP35:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP19:%.*]] = icmp sle i64 [[TMP35]], [[TMP36]]
+// CHECK1-NEXT: [[TMP33:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP19:%.*]] = icmp sle i64 [[TMP33]], [[TMP34]]
// CHECK1-NEXT: br i1 [[CMP19]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP37:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP37]], 1
+// CHECK1-NEXT: [[TMP35:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP35]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
// 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: [[TMP38:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD20:%.*]] = add nsw i64 [[TMP38]], 1
+// CHECK1-NEXT: [[TMP36:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD20:%.*]] = add nsw i64 [[TMP36]], 1
// CHECK1-NEXT: store i64 [[ADD20]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP37:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[TMP37]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP38]])
// CHECK1-NEXT: [[TMP39:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP39]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP40]])
-// CHECK1-NEXT: [[TMP41:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[TMP41]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB2]], i32 [[TMP42]], i32 1)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP45:%.*]] = inttoptr i64 [[TMP5]] to ptr
-// CHECK1-NEXT: store ptr [[TMP45]], ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
-// CHECK1-NEXT: [[TMP48:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP47]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3barv_l55.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP48]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB2]], i32 [[TMP40]], i32 1)
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds [2 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: [[TMP43:%.*]] = inttoptr i64 [[TMP5]] to ptr
+// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[TMP44]], align 4
+// CHECK1-NEXT: [[TMP46:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP45]], i32 1, i64 16, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3barv_l55.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP46]], 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: [[TMP49:%.*]] = getelementptr double, ptr [[ARRAYIDX3]], i64 [[TMP5]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP49]]
+// CHECK1-NEXT: [[TMP47:%.*]] = getelementptr double, ptr [[ARRAYIDX3]], i64 [[TMP5]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP47]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE25:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST21:%.*]] = phi ptr [ [[ARRAYIDX3]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT23:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP50:%.*]] = load double, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST21]], align 8
-// CHECK1-NEXT: [[TMP51:%.*]] = load double, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 8
-// CHECK1-NEXT: [[ADD22:%.*]] = fadd double [[TMP50]], [[TMP51]]
+// CHECK1-NEXT: [[TMP48:%.*]] = load double, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST21]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load double, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 8
+// CHECK1-NEXT: [[ADD22:%.*]] = fadd double [[TMP48]], [[TMP49]]
// CHECK1-NEXT: store double [[ADD22]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST21]], align 8
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT23]] = getelementptr double, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST21]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr double, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE24:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT23]], [[TMP49]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE24:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT23]], [[TMP47]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE24]], label [[OMP_ARRAYCPY_DONE25]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done25:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP47]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB1]], i32 [[TMP45]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP52:%.*]] = getelementptr double, ptr [[ARRAYIDX3]], i64 [[TMP5]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY26:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP52]]
+// CHECK1-NEXT: [[TMP50:%.*]] = getelementptr double, ptr [[ARRAYIDX3]], i64 [[TMP5]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY26:%.*]] = icmp eq ptr [[ARRAYIDX3]], [[TMP50]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY26]], label [[OMP_ARRAYCPY_DONE33:%.*]], label [[OMP_ARRAYCPY_BODY27:%.*]]
// CHECK1: omp.arraycpy.body27:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST28:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT31:%.*]], [[OMP_ARRAYCPY_BODY27]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST29:%.*]] = phi ptr [ [[ARRAYIDX3]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT30:%.*]], [[OMP_ARRAYCPY_BODY27]] ]
-// CHECK1-NEXT: [[TMP53:%.*]] = load double, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST28]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = atomicrmw fadd ptr [[OMP_ARRAYCPY_DESTELEMENTPAST29]], double [[TMP53]] monotonic, align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load double, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST28]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = atomicrmw fadd ptr [[OMP_ARRAYCPY_DESTELEMENTPAST29]], double [[TMP51]] monotonic, align 8
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT30]] = getelementptr double, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST29]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT31]] = getelementptr double, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST28]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE32:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT30]], [[TMP52]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE32:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT30]], [[TMP50]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE32]], label [[OMP_ARRAYCPY_DONE33]], label [[OMP_ARRAYCPY_BODY27]]
// CHECK1: omp.arraycpy.done33:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP55:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP55]])
+// CHECK1-NEXT: [[TMP53:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP53]])
// CHECK1-NEXT: ret void
//
//
diff --git a/clang/test/OpenMP/sections_reduction_task_codegen.cpp b/clang/test/OpenMP/sections_reduction_task_codegen.cpp
index 83145d570e471..0fc426ba5c25b 100644
--- a/clang/test/OpenMP/sections_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/sections_reduction_task_codegen.cpp
@@ -74,8 +74,8 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT: store ptr [[ARGV]], ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
-// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[ARGC_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
+// CHECK1-NEXT: [[TMP1:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_SECTIONS_LB_]], align 4
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_SECTIONS_UB_]], align 4
// CHECK1-NEXT: store i32 1, ptr [[DOTOMP_SECTIONS_ST_]], align 4
@@ -96,217 +96,215 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP9:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP10:%.*]] = sub i64 [[TMP8]], [[TMP9]]
// CHECK1-NEXT: [[TMP11:%.*]] = add nuw i64 [[TMP10]], 1
-// CHECK1-NEXT: [[TMP12:%.*]] = mul nuw i64 [[TMP11]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP13:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP13]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP11]], align 16
// CHECK1-NEXT: store i64 [[TMP11]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP11]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP14]]
+// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP14]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP1]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP16]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP19]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP1]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP20]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP26]], align 8
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP27]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[TMP1]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP30]], i64 0
-// CHECK1-NEXT: [[TMP31:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP32:%.*]] = sext i32 [[TMP31]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP32]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP1]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP33]], i64 9
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP34]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP28]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP38:%.*]] = sub i64 [[TMP36]], [[TMP37]]
-// CHECK1-NEXT: [[TMP39:%.*]] = add nuw i64 [[TMP38]], 1
-// CHECK1-NEXT: [[TMP40:%.*]] = mul nuw i64 [[TMP39]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP40]], ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP45]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
-// CHECK1-NEXT: [[TMP48:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP47]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP48]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = load i32, ptr [[TMP49]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2:[0-9]+]], i32 [[TMP50]], i32 34, ptr [[DOTOMP_SECTIONS_IL_]], ptr [[DOTOMP_SECTIONS_LB_]], ptr [[DOTOMP_SECTIONS_UB_]], ptr [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
-// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
-// CHECK1-NEXT: [[TMP52:%.*]] = icmp slt i32 [[TMP51]], 0
-// CHECK1-NEXT: [[TMP53:%.*]] = select i1 [[TMP52]], i32 [[TMP51]], i32 0
-// CHECK1-NEXT: store i32 [[TMP53]], ptr [[DOTOMP_SECTIONS_UB_]], align 4
-// CHECK1-NEXT: [[TMP54:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_LB_]], align 4
-// CHECK1-NEXT: store i32 [[TMP54]], ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[TMP1]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
+// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP1]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
+// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP38]], ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP43]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP45:%.*]] = load i32, ptr [[TMP44]], align 4
+// CHECK1-NEXT: [[TMP46:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP45]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP46]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = load i32, ptr [[TMP47]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_4(ptr @[[GLOB2:[0-9]+]], i32 [[TMP48]], i32 34, ptr [[DOTOMP_SECTIONS_IL_]], ptr [[DOTOMP_SECTIONS_LB_]], ptr [[DOTOMP_SECTIONS_UB_]], ptr [[DOTOMP_SECTIONS_ST_]], i32 1, i32 1)
+// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
+// CHECK1-NEXT: [[TMP50:%.*]] = icmp slt i32 [[TMP49]], 0
+// CHECK1-NEXT: [[TMP51:%.*]] = select i1 [[TMP50]], i32 [[TMP49]], i32 0
+// CHECK1-NEXT: store i32 [[TMP51]], ptr [[DOTOMP_SECTIONS_UB_]], align 4
+// CHECK1-NEXT: [[TMP52:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_LB_]], align 4
+// CHECK1-NEXT: store i32 [[TMP52]], ptr [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
-// CHECK1-NEXT: [[TMP56:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sle i32 [[TMP55]], [[TMP56]]
+// CHECK1-NEXT: [[TMP53:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: [[TMP54:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_UB_]], align 4
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sle i32 [[TMP53]], [[TMP54]]
// CHECK1-NEXT: br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP57:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
-// CHECK1-NEXT: switch i32 [[TMP57]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
+// CHECK1-NEXT: [[TMP55:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: switch i32 [[TMP55]], label [[DOTOMP_SECTIONS_EXIT:%.*]] [
// CHECK1-NEXT: i32 0, label [[DOTOMP_SECTIONS_CASE:%.*]]
// CHECK1-NEXT: ]
// CHECK1: .omp.sections.case:
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP61]], ptr [[TMP60]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: [[TMP64:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP63]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP64]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP65]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[TMP66]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP67]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP64]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP69:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP68]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP70]], ptr [[TMP69]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[TMP71]], align 4
-// CHECK1-NEXT: [[TMP73:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP72]], ptr [[TMP64]])
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: store ptr [[TMP59]], ptr [[TMP58]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
+// CHECK1-NEXT: [[TMP62:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP61]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP62]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP63]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[TMP64]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP65]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP62]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP66]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP68]], ptr [[TMP67]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[TMP69]], align 4
+// CHECK1-NEXT: [[TMP71:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP70]], ptr [[TMP62]])
// CHECK1-NEXT: br label [[DOTOMP_SECTIONS_EXIT]]
// CHECK1: .omp.sections.exit:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1: omp.inner.for.inc:
-// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
-// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP74]], 1
+// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[DOTOMP_SECTIONS_IV_]], align 4
+// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP72]], 1
// CHECK1-NEXT: store i32 [[INC]], ptr [[DOTOMP_SECTIONS_IV_]], align 4
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND]]
// CHECK1: omp.inner.for.end:
+// CHECK1-NEXT: [[TMP73:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[TMP73]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP74]])
// CHECK1-NEXT: [[TMP75:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[TMP75]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP76]])
-// CHECK1-NEXT: [[TMP77:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[TMP77]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP78]], i32 1)
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP80]], align 8
-// CHECK1-NEXT: [[TMP81:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP82:%.*]] = inttoptr i64 [[TMP11]] to ptr
-// CHECK1-NEXT: store ptr [[TMP82]], ptr [[TMP81]], align 8
-// CHECK1-NEXT: [[TMP83:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP83]], align 4
-// CHECK1-NEXT: [[TMP85:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB3:[0-9]+]], i32 [[TMP84]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP85]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP76]], i32 1)
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP78]], align 8
+// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP80:%.*]] = inttoptr i64 [[TMP11]] to ptr
+// CHECK1-NEXT: store ptr [[TMP80]], ptr [[TMP79]], align 8
+// CHECK1-NEXT: [[TMP81:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP81]], align 4
+// CHECK1-NEXT: [[TMP83:%.*]] = call i32 @__kmpc_reduce(ptr @[[GLOB3:[0-9]+]], i32 [[TMP82]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @main.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP83]], 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: [[TMP86:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP87:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP86]], [[TMP87]]
+// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP85:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP84]], [[TMP85]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP88:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP88]]
+// CHECK1-NEXT: [[TMP86:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE18:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST12:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT16:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP89]] to i32
-// CHECK1-NEXT: [[TMP90:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP90]] to i32
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP87]] to i32
+// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP88]] to i32
// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[CONV]], [[CONV13]]
// CHECK1-NEXT: [[CONV15:%.*]] = trunc i32 [[ADD14]] to i8
// CHECK1-NEXT: store i8 [[CONV15]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT16]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP88]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE17]], label [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done18:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP84]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP91:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP92:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP91]] monotonic, align 4
-// CHECK1-NEXT: [[TMP93:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP93]]
+// CHECK1-NEXT: [[TMP89:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP90:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP89]] monotonic, align 4
+// CHECK1-NEXT: [[TMP91:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP11]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY19]], label [[OMP_ARRAYCPY_DONE32:%.*]], label [[OMP_ARRAYCPY_BODY20:%.*]]
// CHECK1: omp.arraycpy.body20:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST21:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT30:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST22:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT29:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP92]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP95:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP100:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP95]], ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP96]] to i32
-// CHECK1-NEXT: [[TMP97:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP97]] to i32
+// CHECK1-NEXT: [[TMP93:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP98:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP93]], ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP95]] to i32
// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[CONV25]], [[CONV26]]
// CHECK1-NEXT: [[CONV28:%.*]] = trunc i32 [[ADD27]] to i8
// CHECK1-NEXT: store i8 [[CONV28]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP98:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP99:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP95]], i8 [[TMP98]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP100]] = extractvalue { i8, i1 } [[TMP99]], 0
-// CHECK1-NEXT: [[TMP101:%.*]] = extractvalue { i8, i1 } [[TMP99]], 1
-// CHECK1-NEXT: br i1 [[TMP101]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP97:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP93]], i8 [[TMP96]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP98]] = extractvalue { i8, i1 } [[TMP97]], 0
+// CHECK1-NEXT: [[TMP99:%.*]] = extractvalue { i8, i1 } [[TMP97]], 1
+// CHECK1-NEXT: br i1 [[TMP99]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT29]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT30]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP93]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE31]], label [[OMP_ARRAYCPY_DONE32]], label [[OMP_ARRAYCPY_BODY20]]
// CHECK1: omp.arraycpy.done32:
-// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP84]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce(ptr @[[GLOB3]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP102:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP102]])
-// CHECK1-NEXT: [[TMP103:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP104:%.*]] = load i32, ptr [[TMP103]], align 4
-// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB4:[0-9]+]], i32 [[TMP104]])
+// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP100]])
+// CHECK1-NEXT: [[TMP101:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP102:%.*]] = load i32, ptr [[TMP101]], align 4
+// CHECK1-NEXT: call void @__kmpc_barrier(ptr @[[GLOB4:[0-9]+]], i32 [[TMP102]])
// CHECK1-NEXT: ret void
//
//
@@ -430,31 +428,31 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META15:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]]
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META5:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META8:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP10:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP11:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: call void [[TMP10]](ptr [[TMP11]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]]) #[[ATTR6:[0-9]+]], !inline_history [[META15:![0-9]+]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META14]]
// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META14]]
// CHECK1-NEXT: [[TMP17:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP15]], ptr [[TMP14]])
// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[TMP18]], align 8
// CHECK1-NEXT: [[TMP20:%.*]] = load ptr, ptr [[TMP19]], align 8
// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[TMP21]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP23:%.*]] = load i32, ptr [[TMP22]], align 4
// CHECK1-NEXT: [[TMP24:%.*]] = sext i32 [[TMP23]] to i64
// CHECK1-NEXT: [[LB_ADD_LEN_I:%.*]] = add nsw i64 -1, [[TMP24]]
@@ -467,19 +465,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META15]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META14]]
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
-// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META15]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
+// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META14]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META14]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/target_in_reduction_codegen.cpp b/clang/test/OpenMP/target_in_reduction_codegen.cpp
index 3c85936b637e3..fe1068782f5c7 100644
--- a/clang/test/OpenMP/target_in_reduction_codegen.cpp
+++ b/clang/test/OpenMP/target_in_reduction_codegen.cpp
@@ -139,7 +139,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP48]], align 8
// CHECK1-NEXT: [[TMP50:%.*]] = mul nuw i64 [[TMP2]], 2
-// CHECK1-NEXT: [[TMP51:%.*]] = udiv exact i64 [[TMP50]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP51:%.*]] = udiv exact i64 [[TMP50]], 2
// CHECK1-NEXT: [[TMP52:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT: store i64 [[TMP50]], ptr [[TMP52]], align 8
// CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
diff --git a/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
index 49555bfc6ed16..972dff1d4cd4a 100644
--- a/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
@@ -112,223 +112,221 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP6]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP6]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP6]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_7:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
-// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
-// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN10]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP49]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
+// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
+// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN10]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP47]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP48:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP48]], 9
// 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: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP49]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP50]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP51]], [[TMP52]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP55]], 1
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP53]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP56]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP57]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP59]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
-// CHECK1-NEXT: [[TMP62:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP61]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP62]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP63]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[TMP64]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP65]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP62]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP66]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP68]], ptr [[TMP67]], align 8
-// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[TMP69]], align 4
-// CHECK1-NEXT: [[TMP71:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP70]], ptr [[TMP62]])
+// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP54]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP55]], align 8
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP57:%.*]] = load ptr, ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP57]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP58:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load i32, ptr [[TMP58]], align 4
+// CHECK1-NEXT: [[TMP60:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP59]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP61:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP60]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP61]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP63:%.*]] = load ptr, ptr [[TMP62]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP63]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP60]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP64]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP66]], ptr [[TMP65]], align 8
+// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP68:%.*]] = load i32, ptr [[TMP67]], align 4
+// CHECK1-NEXT: [[TMP69:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP68]], ptr [[TMP60]])
// 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: [[TMP72:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP72]], 1
+// CHECK1-NEXT: [[TMP70:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP70]], 1
// CHECK1-NEXT: store i64 [[ADD14]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP71:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[TMP71]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP72]])
// CHECK1-NEXT: [[TMP73:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[TMP73]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP74]])
-// CHECK1-NEXT: [[TMP75:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[TMP75]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP76]], i32 1)
-// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP77]], align 8
-// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP78]], align 8
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP80:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP80]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP81:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP81]], align 4
-// CHECK1-NEXT: [[TMP83:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3:[0-9]+]], i32 [[TMP82]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP83]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP74]], i32 1)
+// CHECK1-NEXT: [[TMP75:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP75]], align 8
+// CHECK1-NEXT: [[TMP76:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP76]], align 8
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP78:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP78]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP79:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP80:%.*]] = load i32, ptr [[TMP79]], align 4
+// CHECK1-NEXT: [[TMP81:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB3:[0-9]+]], i32 [[TMP80]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP81]], 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: [[TMP84:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP85:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP84]], [[TMP85]]
+// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP83:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP82]], [[TMP83]]
// CHECK1-NEXT: store i32 [[ADD15]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP86:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP86]]
+// CHECK1-NEXT: [[TMP84:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP84]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE22:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST16:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT20:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP87]] to i32
-// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP88]] to i32
+// CHECK1-NEXT: [[TMP85:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP85]] to i32
+// CHECK1-NEXT: [[TMP86:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP86]] to i32
// CHECK1-NEXT: [[ADD18:%.*]] = add nsw i32 [[CONV]], [[CONV17]]
// CHECK1-NEXT: [[CONV19:%.*]] = trunc i32 [[ADD18]] to i8
// CHECK1-NEXT: store i8 [[CONV19]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT20]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP86]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP84]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_DONE22]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done22:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB3]], i32 [[TMP80]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP89:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP90:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP89]] monotonic, align 4
-// CHECK1-NEXT: [[TMP91:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP91]]
+// CHECK1-NEXT: [[TMP87:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP88:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP87]] monotonic, align 4
+// CHECK1-NEXT: [[TMP89:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP89]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY23]], label [[OMP_ARRAYCPY_DONE36:%.*]], label [[OMP_ARRAYCPY_BODY24:%.*]]
// CHECK1: omp.arraycpy.body24:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST25:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT34:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT33:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP92]] to i32
+// CHECK1-NEXT: [[TMP90:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP90]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP93:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP98:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP93]], ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP94]] to i32
-// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP95]] to i32
+// CHECK1-NEXT: [[TMP91:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP96:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP91]], ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP92]] to i32
+// CHECK1-NEXT: [[TMP93:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP93]] to i32
// CHECK1-NEXT: [[ADD31:%.*]] = add nsw i32 [[CONV29]], [[CONV30]]
// CHECK1-NEXT: [[CONV32:%.*]] = trunc i32 [[ADD31]] to i8
// CHECK1-NEXT: store i8 [[CONV32]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP97:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP93]], i8 [[TMP96]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP98]] = extractvalue { i8, i1 } [[TMP97]], 0
-// CHECK1-NEXT: [[TMP99:%.*]] = extractvalue { i8, i1 } [[TMP97]], 1
-// CHECK1-NEXT: br i1 [[TMP99]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP95:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP91]], i8 [[TMP94]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP96]] = extractvalue { i8, i1 } [[TMP95]], 0
+// CHECK1-NEXT: [[TMP97:%.*]] = extractvalue { i8, i1 } [[TMP95]], 1
+// CHECK1-NEXT: br i1 [[TMP97]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT34]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP91]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP89]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_DONE36]], label [[OMP_ARRAYCPY_BODY24]]
// CHECK1: omp.arraycpy.done36:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP100]])
+// CHECK1-NEXT: [[TMP98:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP98]])
// CHECK1-NEXT: ret void
//
//
@@ -489,19 +487,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp b/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
index 8ecb6e4412245..bebb11fe0d2e6 100644
--- a/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
@@ -101,180 +101,178 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 0, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP48]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP49]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP51:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP51]], ptr [[TMP50]], align 8
-// CHECK1-NEXT: [[TMP52:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP53:%.*]] = load i32, ptr [[TMP52]], align 4
-// CHECK1-NEXT: [[TMP54:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP53]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP55:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP54]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP55]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP57:%.*]] = load ptr, ptr [[TMP56]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP57]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP54]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP58]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP60]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load i32, ptr [[TMP61]], align 4
-// CHECK1-NEXT: [[TMP63:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP62]], ptr [[TMP54]])
-// CHECK1-NEXT: [[TMP64:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP65:%.*]] = load i32, ptr [[TMP64]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP65]], i32 0)
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP66]], align 8
-// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP67]], align 8
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP69:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP69]], ptr [[TMP68]], align 8
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[TMP70]], align 4
-// CHECK1-NEXT: [[TMP72:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP71]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP72]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 0, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP46]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP47]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP49:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: store ptr [[TMP49]], ptr [[TMP48]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[TMP50]], align 4
+// CHECK1-NEXT: [[TMP52:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP51]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP53:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP52]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP54:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP53]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP55:%.*]] = load ptr, ptr [[TMP54]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP55]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP52]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP56]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP58:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP58]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load i32, ptr [[TMP59]], align 4
+// CHECK1-NEXT: [[TMP61:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP60]], ptr [[TMP52]])
+// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP63]], i32 0)
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP64]], align 8
+// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP65]], align 8
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP67:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP67]], ptr [[TMP66]], align 8
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
+// CHECK1-NEXT: [[TMP70:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB2:[0-9]+]], i32 [[TMP69]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP70]], 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: [[TMP73:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP73]], [[TMP74]]
+// CHECK1-NEXT: [[TMP71:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP71]], [[TMP72]]
// CHECK1-NEXT: store i32 [[ADD]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP75:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP75]]
+// CHECK1-NEXT: [[TMP73:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP73]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE18:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST12:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT16:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP76:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP76]] to i32
-// CHECK1-NEXT: [[TMP77:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP77]] to i32
+// CHECK1-NEXT: [[TMP74:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP74]] to i32
+// CHECK1-NEXT: [[TMP75:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV13:%.*]] = sext i8 [[TMP75]] to i32
// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[CONV]], [[CONV13]]
// CHECK1-NEXT: [[CONV15:%.*]] = trunc i32 [[ADD14]] to i8
// CHECK1-NEXT: store i8 [[CONV15]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT16]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST12]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP75]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE17:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT16]], [[TMP73]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE17]], label [[OMP_ARRAYCPY_DONE18]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done18:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP71]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB2]], i32 [[TMP69]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP79:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP78]] monotonic, align 4
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP80]]
+// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP77:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP76]] monotonic, align 4
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY19:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP78]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY19]], label [[OMP_ARRAYCPY_DONE32:%.*]], label [[OMP_ARRAYCPY_BODY20:%.*]]
// CHECK1: omp.arraycpy.body20:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST21:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT30:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST22:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT29:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP81]] to i32
+// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV23:%.*]] = sext i8 [[TMP79]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP82:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP87:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP82]], ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[_TMP24]], align 1
-// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP83]] to i32
-// CHECK1-NEXT: [[TMP84:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP84]] to i32
+// CHECK1-NEXT: [[TMP80:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY20]] ], [ [[TMP85:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP80]], ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[_TMP24]], align 1
+// CHECK1-NEXT: [[CONV25:%.*]] = sext i8 [[TMP81]] to i32
+// CHECK1-NEXT: [[TMP82:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP82]] to i32
// CHECK1-NEXT: [[ADD27:%.*]] = add nsw i32 [[CONV25]], [[CONV26]]
// CHECK1-NEXT: [[CONV28:%.*]] = trunc i32 [[ADD27]] to i8
// CHECK1-NEXT: store i8 [[CONV28]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP85:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP86:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP82]], i8 [[TMP85]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP87]] = extractvalue { i8, i1 } [[TMP86]], 0
-// CHECK1-NEXT: [[TMP88:%.*]] = extractvalue { i8, i1 } [[TMP86]], 1
-// CHECK1-NEXT: br i1 [[TMP88]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP84:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i8 [[TMP80]], i8 [[TMP83]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP85]] = extractvalue { i8, i1 } [[TMP84]], 0
+// CHECK1-NEXT: [[TMP86:%.*]] = extractvalue { i8, i1 } [[TMP84]], 1
+// CHECK1-NEXT: br i1 [[TMP86]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT29]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST22]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT30]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST21]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP80]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE31:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT29]], [[TMP78]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE31]], label [[OMP_ARRAYCPY_DONE32]], label [[OMP_ARRAYCPY_BODY20]]
// CHECK1: omp.arraycpy.done32:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP89:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP89]])
+// CHECK1-NEXT: [[TMP87:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP87]])
// CHECK1-NEXT: ret void
//
//
@@ -435,19 +433,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
index d366b6530806c..ff113accabd12 100644
--- a/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
@@ -107,203 +107,201 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
// CHECK1-NEXT: store i64 0, ptr [[DOTOMP_COMB_LB]], align 8
// CHECK1-NEXT: store i64 9, ptr [[DOTOMP_COMB_UB]], align 8
// CHECK1-NEXT: store i64 1, ptr [[DOTOMP_STRIDE]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
-// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP49]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP47]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP48:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP48]], 9
// 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: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP49]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP50]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP51]], [[TMP52]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp_outlined, i64 [[TMP55]], i64 [[TMP56]], ptr [[ARGC1]], ptr [[TMP57]])
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp_outlined, i64 [[TMP53]], i64 [[TMP54]], ptr [[ARGC1]], ptr [[TMP55]])
// CHECK1-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1: omp.inner.for.inc:
-// CHECK1-NEXT: [[TMP58:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = load i64, ptr [[DOTOMP_STRIDE]], align 8
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 [[TMP58]], [[TMP59]]
+// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = load i64, ptr [[DOTOMP_STRIDE]], align 8
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 [[TMP56]], [[TMP57]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP58:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load i32, ptr [[TMP58]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP59]])
// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP61]])
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP63]], i32 1)
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP64]], align 8
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP65]], align 8
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP67:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP67]], ptr [[TMP66]], align 8
-// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
-// CHECK1-NEXT: [[TMP70:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4:[0-9]+]], i32 [[TMP69]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP70]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP61]], i32 1)
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP62]], align 8
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP63]], align 8
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP65:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP65]], ptr [[TMP64]], align 8
+// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP67:%.*]] = load i32, ptr [[TMP66]], align 4
+// CHECK1-NEXT: [[TMP68:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4:[0-9]+]], i32 [[TMP67]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP68]], 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: [[TMP71:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[TMP71]], [[TMP72]]
+// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[TMP69]], [[TMP70]]
// CHECK1-NEXT: store i32 [[ADD14]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP73:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP73]]
+// CHECK1-NEXT: [[TMP71:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP71]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE21:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST15:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT19:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP74:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP74]] to i32
-// CHECK1-NEXT: [[TMP75:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV16:%.*]] = sext i8 [[TMP75]] to i32
+// CHECK1-NEXT: [[TMP72:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP72]] to i32
+// CHECK1-NEXT: [[TMP73:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV16:%.*]] = sext i8 [[TMP73]] to i32
// CHECK1-NEXT: [[ADD17:%.*]] = add nsw i32 [[CONV]], [[CONV16]]
// CHECK1-NEXT: [[CONV18:%.*]] = trunc i32 [[ADD17]] to i8
// CHECK1-NEXT: store i8 [[CONV18]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT19]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT19]], [[TMP73]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT19]], [[TMP71]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done21:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP69]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP67]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP77:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP76]] monotonic, align 4
-// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY22:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP78]]
+// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP75:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP74]] monotonic, align 4
+// CHECK1-NEXT: [[TMP76:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY22:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP76]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY22]], label [[OMP_ARRAYCPY_DONE35:%.*]], label [[OMP_ARRAYCPY_BODY23:%.*]]
// CHECK1: omp.arraycpy.body23:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST24:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT33:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST25:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT32:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP79]] to i32
+// CHECK1-NEXT: [[TMP77:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP77]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP80:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY23]] ], [ [[TMP85:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP80]], ptr [[_TMP27]], align 1
-// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[_TMP27]], align 1
-// CHECK1-NEXT: [[CONV28:%.*]] = sext i8 [[TMP81]] to i32
-// CHECK1-NEXT: [[TMP82:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP82]] to i32
+// CHECK1-NEXT: [[TMP78:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY23]] ], [ [[TMP83:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP78]], ptr [[_TMP27]], align 1
+// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[_TMP27]], align 1
+// CHECK1-NEXT: [[CONV28:%.*]] = sext i8 [[TMP79]] to i32
+// CHECK1-NEXT: [[TMP80:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP80]] to i32
// CHECK1-NEXT: [[ADD30:%.*]] = add nsw i32 [[CONV28]], [[CONV29]]
// CHECK1-NEXT: [[CONV31:%.*]] = trunc i32 [[ADD30]] to i8
// CHECK1-NEXT: store i8 [[CONV31]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP84:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]], i8 [[TMP80]], i8 [[TMP83]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP85]] = extractvalue { i8, i1 } [[TMP84]], 0
-// CHECK1-NEXT: [[TMP86:%.*]] = extractvalue { i8, i1 } [[TMP84]], 1
-// CHECK1-NEXT: br i1 [[TMP86]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP82:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]], i8 [[TMP78]], i8 [[TMP81]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP83]] = extractvalue { i8, i1 } [[TMP82]], 0
+// CHECK1-NEXT: [[TMP84:%.*]] = extractvalue { i8, i1 } [[TMP82]], 1
+// CHECK1-NEXT: br i1 [[TMP84]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT32]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE34:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT32]], [[TMP78]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE34:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT32]], [[TMP76]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE34]], label [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_BODY23]]
// CHECK1: omp.arraycpy.done35:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP87:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP87]])
+// CHECK1-NEXT: [[TMP85:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP85]])
// CHECK1-NEXT: ret void
//
//
@@ -448,223 +446,221 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP11:%.*]] = sub i64 [[TMP9]], [[TMP10]]
// CHECK1-NEXT: [[TMP12:%.*]] = add nuw i64 [[TMP11]], 1
-// CHECK1-NEXT: [[TMP13:%.*]] = mul nuw i64 [[TMP12]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP14:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP14]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP13:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP13]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP12]], align 16
// CHECK1-NEXT: store i64 [[TMP12]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP15]]
+// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP14]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP15]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP14]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP17:%.*]] = load ptr, ptr [[TMP16]], align 8
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[TMP17]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP20:%.*]] = sub i64 [[TMP18]], [[TMP19]]
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP20]]
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP16]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP19]]
// CHECK1-NEXT: store ptr [[_TMP6]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP21]], ptr [[_TMP6]], align 8
+// CHECK1-NEXT: store ptr [[TMP20]], ptr [[_TMP6]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T_0]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..3, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP26]], align 8
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..4, ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP28]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..3, ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..4, ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP27]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_7:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T_0]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP30]], i64 0
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
-// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP31]], i64 0
-// CHECK1-NEXT: [[TMP32:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP33:%.*]] = sext i32 [[TMP32]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP33]]
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP34]], i64 9
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
-// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP35]], i64 [[LB_ADD_LEN10]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP29]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
-// CHECK1-NEXT: [[TMP40:%.*]] = add nuw i64 [[TMP39]], 1
-// CHECK1-NEXT: [[TMP41:%.*]] = mul nuw i64 [[TMP40]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP41]], ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..5, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..6, ptr [[TMP45]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP46]], align 8
-// CHECK1-NEXT: [[TMP47:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load i32, ptr [[TMP47]], align 4
-// CHECK1-NEXT: [[TMP49:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP48]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP49]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[TMP50]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP51]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP52]], 9
+// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP29]], i64 0
+// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
+// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP30]], i64 0
+// CHECK1-NEXT: [[TMP31:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP32:%.*]] = sext i32 [[TMP31]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP32]]
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP33]], i64 9
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
+// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP34]], i64 [[LB_ADD_LEN10]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP28]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP35]], align 8
+// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = sub i64 [[TMP36]], [[TMP37]]
+// CHECK1-NEXT: [[TMP39:%.*]] = add nuw i64 [[TMP38]], 1
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..5, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..6, ptr [[TMP43]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
+// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
+// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP49]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
// 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: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP53]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP54]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP55]], [[TMP56]]
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP57:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP57]], 1
+// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP55]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP61]], ptr [[TMP60]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: [[TMP64:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP63]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP64]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP65]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[TMP66]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP67]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP64]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP69:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP68]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP70]], ptr [[TMP69]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[TMP71]], align 4
-// CHECK1-NEXT: [[TMP73:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP72]], ptr [[TMP64]])
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP59]], ptr [[TMP58]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
+// CHECK1-NEXT: [[TMP62:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP61]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP62]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP63]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[TMP64]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP65]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP62]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP66]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP68]], ptr [[TMP67]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[TMP69]], align 4
+// CHECK1-NEXT: [[TMP71:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP70]], ptr [[TMP62]])
// 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: [[TMP74:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP74]], 1
+// CHECK1-NEXT: [[TMP72:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP72]], 1
// CHECK1-NEXT: store i64 [[ADD14]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP73:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[TMP73]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP74]])
// CHECK1-NEXT: [[TMP75:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[TMP75]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP76]])
-// CHECK1-NEXT: [[TMP77:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[TMP77]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP78]], i32 1)
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP80]], align 8
-// CHECK1-NEXT: [[TMP81:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP82:%.*]] = inttoptr i64 [[TMP12]] to ptr
-// CHECK1-NEXT: store ptr [[TMP82]], ptr [[TMP81]], align 8
-// CHECK1-NEXT: [[TMP83:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP83]], align 4
-// CHECK1-NEXT: [[TMP85:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP84]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP85]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP76]], i32 1)
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP78]], align 8
+// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP80:%.*]] = inttoptr i64 [[TMP12]] to ptr
+// CHECK1-NEXT: store ptr [[TMP80]], ptr [[TMP79]], align 8
+// CHECK1-NEXT: [[TMP81:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP81]], align 4
+// CHECK1-NEXT: [[TMP83:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP82]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l14.omp_outlined.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP83]], 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: [[TMP86:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP87:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP86]], [[TMP87]]
+// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP85:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP84]], [[TMP85]]
// CHECK1-NEXT: store i32 [[ADD15]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP88:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP88]]
+// CHECK1-NEXT: [[TMP86:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE22:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST16:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT20:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP89]] to i32
-// CHECK1-NEXT: [[TMP90:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP90]] to i32
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP87]] to i32
+// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP88]] to i32
// CHECK1-NEXT: [[ADD18:%.*]] = add nsw i32 [[CONV]], [[CONV17]]
// CHECK1-NEXT: [[CONV19:%.*]] = trunc i32 [[ADD18]] to i8
// CHECK1-NEXT: store i8 [[CONV19]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT20]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP88]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_DONE22]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done22:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP84]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP91:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP92:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP91]] monotonic, align 4
-// CHECK1-NEXT: [[TMP93:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP93]]
+// CHECK1-NEXT: [[TMP89:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP90:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP89]] monotonic, align 4
+// CHECK1-NEXT: [[TMP91:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY23]], label [[OMP_ARRAYCPY_DONE36:%.*]], label [[OMP_ARRAYCPY_BODY24:%.*]]
// CHECK1: omp.arraycpy.body24:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST25:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT34:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT33:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP92]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP95:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP100:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP95]], ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP96]] to i32
-// CHECK1-NEXT: [[TMP97:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP97]] to i32
+// CHECK1-NEXT: [[TMP93:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP98:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP93]], ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP95]] to i32
// CHECK1-NEXT: [[ADD31:%.*]] = add nsw i32 [[CONV29]], [[CONV30]]
// CHECK1-NEXT: [[CONV32:%.*]] = trunc i32 [[ADD31]] to i8
// CHECK1-NEXT: store i8 [[CONV32]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP98:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP99:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP95]], i8 [[TMP98]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP100]] = extractvalue { i8, i1 } [[TMP99]], 0
-// CHECK1-NEXT: [[TMP101:%.*]] = extractvalue { i8, i1 } [[TMP99]], 1
-// CHECK1-NEXT: br i1 [[TMP101]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP97:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP93]], i8 [[TMP96]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP98]] = extractvalue { i8, i1 } [[TMP97]], 0
+// CHECK1-NEXT: [[TMP99:%.*]] = extractvalue { i8, i1 } [[TMP97]], 1
+// CHECK1-NEXT: br i1 [[TMP99]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT34]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP93]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_DONE36]], label [[OMP_ARRAYCPY_BODY24]]
// CHECK1: omp.arraycpy.done36:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP102:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP102]])
+// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP100]])
// CHECK1-NEXT: ret void
//
//
@@ -825,19 +821,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/task_in_reduction_codegen.cpp b/clang/test/OpenMP/task_in_reduction_codegen.cpp
index cd6ed42218f38..0ba8b6e504491 100644
--- a/clang/test/OpenMP/task_in_reduction_codegen.cpp
+++ b/clang/test/OpenMP/task_in_reduction_codegen.cpp
@@ -159,7 +159,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP35:%.*]] = mul nuw i64 [[TMP2]], 2
-// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 2
// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP37]], align 8
// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
@@ -381,7 +381,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: call void @_ZN1SC1ERKS_(ptr nonnull align 4 dereferenceable(4) [[AGG_RESULT]], ptr nonnull align 4 dereferenceable(4) [[TMP0]])
// CHECK1-NEXT: ret void
//
@@ -472,11 +472,11 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[D]], ptr [[D_ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_1]], ptr [[DOTTASK_RED__ADDR2]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META3]], !align [[META6:![0-9]+]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META3]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META2]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META2]], !align [[META5]]
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
@@ -546,43 +546,42 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 0
// CHECK1-NEXT: [[TMP7:%.*]] = load ptr, ptr [[TMP6]], align 8
// CHECK1-NEXT: [[TMP8:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP3]], i32 0, i32 1
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META14:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META16:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META16]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META15:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META15]]
// CHECK1-NEXT: [[TMP10:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP9]], i32 0, i32 1
// CHECK1-NEXT: [[TMP11:%.*]] = load i64, ptr [[TMP10]], align 8
-// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: call void [[TMP12]](ptr [[TMP13]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META16]]
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP9]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP12:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: call void [[TMP12]](ptr [[TMP13]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]], !inline_history [[META16:![0-9]+]]
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META15]]
+// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP9]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP17:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP18:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META15]]
// CHECK1-NEXT: [[TMP19:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP18]], ptr [[TMP17]], ptr [[TMP16]])
// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP21:%.*]] = load ptr, ptr [[TMP20]], align 8, !nonnull [[META3]], !align [[META5]]
+// CHECK1-NEXT: [[TMP21:%.*]] = load ptr, ptr [[TMP20]], align 8, !nonnull [[META2]], !align [[META4]]
// CHECK1-NEXT: [[TMP22:%.*]] = mul nuw i64 [[TMP11]], 2
-// CHECK1-NEXT: [[TMP23:%.*]] = udiv exact i64 [[TMP22]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP24:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP18]], ptr @{{reduction_size[.].+[.]}})
-// CHECK1-NEXT: store i64 [[TMP23]], ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[TMP15]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP18]], ptr [[TMP25]], ptr [[TMP21]])
-// CHECK1-NEXT: [[TMP27:%.*]] = load i32, ptr [[TMP19]], align 4
-// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP27]] to i64
-// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP26]], i64 [[IDXPROM_I]]
-// CHECK1-NEXT: [[TMP28:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2
-// CHECK1-NEXT: [[CONV_I:%.*]] = sext i16 [[TMP28]] to i32
-// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP19]], align 4
-// CHECK1-NEXT: [[ADD_I:%.*]] = add nsw i32 [[TMP29]], [[CONV_I]]
+// CHECK1-NEXT: [[TMP23:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP18]], ptr @{{reduction_size[.].+[.]}})
+// CHECK1-NEXT: store i64 [[TMP11]], ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[TMP15]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP18]], ptr [[TMP24]], ptr [[TMP21]])
+// CHECK1-NEXT: [[TMP26:%.*]] = load i32, ptr [[TMP19]], align 4
+// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP26]] to i64
+// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP25]], i64 [[IDXPROM_I]]
+// CHECK1-NEXT: [[TMP27:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2
+// CHECK1-NEXT: [[CONV_I:%.*]] = sext i16 [[TMP27]] to i32
+// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[TMP19]], align 4
+// CHECK1-NEXT: [[ADD_I:%.*]] = add nsw i32 [[TMP28]], [[CONV_I]]
// CHECK1-NEXT: store i32 [[ADD_I]], ptr [[TMP19]], align 4
// CHECK1-NEXT: ret i32 0
//
@@ -617,7 +616,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META28]]
// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META28]]
// CHECK1-NEXT: [[TMP8:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META28]]
-// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[TMP8]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP9:%.*]] = load ptr, ptr [[TMP8]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP10:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META28]]
// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP10]], ptr null, ptr [[TMP9]])
// CHECK1-NEXT: [[TMP12:%.*]] = load i32, ptr [[TMP11]], align 4
diff --git a/clang/test/OpenMP/taskgroup_task_reduction_codegen.cpp b/clang/test/OpenMP/taskgroup_task_reduction_codegen.cpp
index a8577c7a13579..63622818db46b 100644
--- a/clang/test/OpenMP/taskgroup_task_reduction_codegen.cpp
+++ b/clang/test/OpenMP/taskgroup_task_reduction_codegen.cpp
@@ -133,7 +133,7 @@ int main(int argc, char **argv) {
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw [[T2]], ptr [[GEPVLA]], i32 0, i32 1
// CHECK-DAG: [[GEPVLA]] = getelementptr inbounds nuw [2 x [[T2]]], ptr [[RD_IN2]], i64 0, i64
// CHECK-DAG: [[TMP40:%.+]] = mul nuw i64 [[VLA_SIZE]], 2
-// CHECK-DAG: [[TMP41:%.+]] = udiv exact i64 [[TMP40]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP41:%.+]] = udiv exact i64 [[TMP40]], 2
// CHECK-DAG: [[TMP42:%.+]] = getelementptr inbounds nuw [[T2]], ptr [[GEPVLA]], i32 0, i32 2
// CHECK-DAG: store i64 [[TMP40]], ptr [[TMP42]],
// CHECK-DAG: [[TMP43:%.+]] = getelementptr inbounds nuw [[T2]], ptr [[GEPVLA]], i32 0, i32 3
diff --git a/clang/test/OpenMP/taskloop_in_reduction_codegen.cpp b/clang/test/OpenMP/taskloop_in_reduction_codegen.cpp
index 1ea8c597f3608..c6c2931cf7472 100644
--- a/clang/test/OpenMP/taskloop_in_reduction_codegen.cpp
+++ b/clang/test/OpenMP/taskloop_in_reduction_codegen.cpp
@@ -145,7 +145,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP35:%.*]] = mul nuw i64 [[TMP2]], 2
-// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 2
// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP37]], align 8
// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
@@ -359,7 +359,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: call void @_ZN1SC1ERKS_(ptr nonnull align 4 dereferenceable(4) [[AGG_RESULT]], ptr nonnull align 4 dereferenceable(4) [[TMP0]])
// CHECK1-NEXT: ret void
//
@@ -451,11 +451,11 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[D]], ptr [[D_ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_1]], ptr [[DOTTASK_RED__ADDR2]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META3]], !align [[META6:![0-9]+]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META3]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META2]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META2]], !align [[META5]]
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
@@ -553,66 +553,65 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 8
// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 9
// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[TMP17]], align 8
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META14:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META16:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META15:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP19]], i32 0, i32 1
// CHECK1-NEXT: [[TMP21:%.*]] = load i64, ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
-// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]], !inline_history [[META18:![0-9]+]]
+// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: [[TMP29:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP27]], ptr [[TMP26]])
// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP19]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META3]], !align [[META5]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META2]], !align [[META4]]
// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP21]], 2
-// CHECK1-NEXT: [[TMP33:%.*]] = udiv exact i64 [[TMP32]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
-// CHECK1-NEXT: store i64 [[TMP33]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP35]], ptr [[TMP31]])
-// CHECK1-NEXT: [[TMP37:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP37]] to i32
-// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
+// CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP34]], ptr [[TMP31]])
+// CHECK1-NEXT: [[TMP36:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP36]] to i32
+// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1: omp.inner.for.cond.i:
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP39]]
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP38]]
// CHECK1-NEXT: br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__EXIT:%.*]]
// CHECK1: omp.inner.for.body.i:
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: store i32 [[TMP40]], ptr [[I_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[TMP29]], align 4
-// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP41]] to i64
-// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP36]], i64 [[IDXPROM_I]]
-// CHECK1-NEXT: [[TMP42:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2
-// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP42]] to i32
-// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[TMP29]], align 4
-// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP43]], [[CONV3_I]]
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: store i32 [[TMP39]], ptr [[I_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP29]], align 4
+// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP40]] to i64
+// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP35]], i64 [[IDXPROM_I]]
+// CHECK1-NEXT: [[TMP41:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2
+// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP41]] to i32
+// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[TMP29]], align 4
+// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP42]], [[CONV3_I]]
// CHECK1-NEXT: store i32 [[ADD4_I]], ptr [[TMP29]], align 4
-// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP44]], 1
-// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP43]], 1
+// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I]]
// CHECK1: .omp_outlined..exit:
// CHECK1-NEXT: ret i32 0
diff --git a/clang/test/OpenMP/taskloop_reduction_codegen.cpp b/clang/test/OpenMP/taskloop_reduction_codegen.cpp
index acd8d666c5768..39997789b896c 100644
--- a/clang/test/OpenMP/taskloop_reduction_codegen.cpp
+++ b/clang/test/OpenMP/taskloop_reduction_codegen.cpp
@@ -93,10 +93,9 @@ sum = 0.0;
// CHECK-DAG: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK-DAG: [[TMP33:%.*]] = ptrtoaddr ptr [[ARRAYIDX5]] to i64
// CHECK-DAG: [[TMP34:%.*]] = sub i64 [[TMP32]], [[TMP33]]
-// CHECK-DAG: [[TMP35:%.*]] = sdiv exact i64 [[TMP34]], 4
-// CHECK-DAG: [[TMP36:%.*]] = add nuw i64 [[TMP35]], 1
-// CHECK-DAG: [[TMP37:%.*]] = mul nuw i64 [[TMP36]], ptrtoint (ptr getelementptr (%struct.S, ptr null, i32 1) to i64)
-// CHECK-DAG: store i64 [[TMP37]], ptr [[TMP38:%[^,]+]],
+// CHECK-DAG: [[TMP35:%.*]] = add nuw i64 [[TMP34]], 4
+// CHECK-DAG: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 4
+// CHECK-DAG: store i64 [[TMP35]], ptr [[TMP38:%[^,]+]],
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK-DAG: [[TMP39:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK-DAG: store ptr @[[RED_INIT2:.+]], ptr [[TMP39]],
@@ -126,7 +125,7 @@ sum = 0.0;
// CHECK-DAG: store ptr [[VLA]], ptr [[TMP52:%[^,]+]],
// CHECK-DAG: [[TMP52]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 1
// CHECK-DAG: [[TMP54:%.*]] = mul nuw i64 [[TMP2]], 4
-// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], ptrtoint (ptr getelementptr (float, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], 4
// CHECK-DAG: store i64 [[TMP54]], ptr [[TMP56:%[^,]+]],
// CHECK-DAG: [[TMP56]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 2
// CHECK-DAG: [[TMP57:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 3
diff --git a/clang/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp b/clang/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp
index b3d32f9a11d3e..2670de75aa356 100644
--- a/clang/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp
+++ b/clang/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp
@@ -145,7 +145,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 1
// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP35:%.*]] = mul nuw i64 [[TMP2]], 2
-// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
+// CHECK1-NEXT: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 2
// CHECK1-NEXT: [[TMP37:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 2
// CHECK1-NEXT: store i64 [[TMP35]], ptr [[TMP37]], align 8
// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_5]], i32 0, i32 3
@@ -359,7 +359,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 8
// CHECK1-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT: store ptr [[B]], ptr [[B_ADDR]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3:![0-9]+]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2:![0-9]+]], !align [[META3:![0-9]+]]
// CHECK1-NEXT: call void @_ZN1SC1ERKS_(ptr nonnull align 4 dereferenceable(4) [[AGG_RESULT]], ptr nonnull align 4 dereferenceable(4) [[TMP0]])
// CHECK1-NEXT: ret void
//
@@ -451,11 +451,11 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: store ptr [[D]], ptr [[D_ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[DOTTASK_RED__ADDR]], align 8
// CHECK1-NEXT: store ptr [[DOTTASK_RED_1]], ptr [[DOTTASK_RED__ADDR2]], align 8
-// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR]], align 8
-// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META3]], !align [[META5:![0-9]+]]
-// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META3]], !align [[META6:![0-9]+]]
-// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META3]], !align [[META6]]
+// CHECK1-NEXT: [[TMP2:%.*]] = load ptr, ptr [[D_ADDR]], align 8, !nonnull [[META2]], !align [[META4:![0-9]+]]
+// CHECK1-NEXT: [[TMP3:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR]], align 8, !nonnull [[META2]], !align [[META5:![0-9]+]]
+// CHECK1-NEXT: [[TMP4:%.*]] = load ptr, ptr [[DOTTASK_RED__ADDR2]], align 8, !nonnull [[META2]], !align [[META5]]
// CHECK1-NEXT: [[TMP5:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP5]], align 8
// CHECK1-NEXT: [[TMP6:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
@@ -553,66 +553,65 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 8
// CHECK1-NEXT: [[TMP17:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T]], ptr [[TMP4]], i32 0, i32 9
// CHECK1-NEXT: [[TMP18:%.*]] = load ptr, ptr [[TMP17]], align 8
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META7:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META10:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META12:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META14:![0-9]+]])
-// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META16:![0-9]+]])
-// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18:![0-9]+]]
-// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META18]]
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META6:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META9:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META11:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META13:![0-9]+]])
+// CHECK1-NEXT: call void @llvm.experimental.noalias.scope.decl(metadata [[META15:![0-9]+]])
+// CHECK1-NEXT: store i32 [[TMP2]], ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17:![0-9]+]]
+// CHECK1-NEXT: store ptr [[TMP5]], ptr [[DOTPART_ID__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP8]], ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr @.omp_task_privates_map., ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP3]], ptr [[DOTTASK_T__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP10]], ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP12]], ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i64 [[TMP14]], ptr [[DOTST__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store i32 [[TMP16]], ptr [[DOTLITER__ADDR_I]], align 4, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[DOTREDUCTIONS__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: store ptr [[TMP7]], ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP19:%.*]] = load ptr, ptr [[__CONTEXT_ADDR_I]], align 8, !noalias [[META17]]
// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON:%.*]], ptr [[TMP19]], i32 0, i32 1
// CHECK1-NEXT: [[TMP21:%.*]] = load i64, ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]]
-// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META3]], !align [[META4]]
+// CHECK1-NEXT: [[TMP22:%.*]] = load ptr, ptr [[DOTCOPY_FN__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP23:%.*]] = load ptr, ptr [[DOTPRIVATES__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: call void [[TMP22]](ptr [[TMP23]], ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]]) #[[ATTR3]], !inline_history [[META18:![0-9]+]]
+// CHECK1-NEXT: [[TMP24:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP25:%.*]] = load ptr, ptr [[DOTFIRSTPRIV_PTR_ADDR1_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[TMP26:%.*]] = load ptr, ptr [[TMP19]], align 8, !nonnull [[META2]], !align [[META3]]
// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP28:%.*]] = load i32, ptr [[DOTGLOBAL_TID__ADDR_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: [[TMP29:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP27]], ptr [[TMP26]])
// CHECK1-NEXT: [[TMP30:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP19]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META3]], !align [[META5]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[TMP30]], align 8, !nonnull [[META2]], !align [[META4]]
// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP21]], 2
-// CHECK1-NEXT: [[TMP33:%.*]] = udiv exact i64 [[TMP32]], ptrtoint (ptr getelementptr (i16, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
-// CHECK1-NEXT: store i64 [[TMP33]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP35]], ptr [[TMP31]])
-// CHECK1-NEXT: [[TMP37:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META18]]
-// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP37]] to i32
-// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]]
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_threadprivate_cached(ptr @[[GLOB1]], i32 [[TMP28]], ptr @{{reduction_size[.].+[.]}})
+// CHECK1-NEXT: store i64 [[TMP21]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP28]], ptr [[TMP34]], ptr [[TMP31]])
+// CHECK1-NEXT: [[TMP36:%.*]] = load i64, ptr [[DOTLB__ADDR_I]], align 8, !noalias [[META17]]
+// CHECK1-NEXT: [[CONV_I:%.*]] = trunc i64 [[TMP36]] to i32
+// CHECK1-NEXT: store i32 [[CONV_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I:%.*]]
// CHECK1: omp.inner.for.cond.i:
-// CHECK1-NEXT: [[TMP38:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19:![0-9]+]]
-// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP39]]
+// CHECK1-NEXT: [[TMP37:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19:![0-9]+]]
+// CHECK1-NEXT: [[CONV2_I:%.*]] = sext i32 [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = load i64, ptr [[DOTUB__ADDR_I]], align 8, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[CMP_I:%.*]] = icmp ule i64 [[CONV2_I]], [[TMP38]]
// CHECK1-NEXT: br i1 [[CMP_I]], label [[OMP_INNER_FOR_BODY_I:%.*]], label [[DOTOMP_OUTLINED__EXIT:%.*]]
// CHECK1: omp.inner.for.body.i:
-// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: store i32 [[TMP40]], ptr [[I_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[TMP41:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP41]] to i64
-// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP36]], i64 [[IDXPROM_I]]
-// CHECK1-NEXT: [[TMP42:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP42]] to i32
-// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP43]], [[CONV3_I]]
+// CHECK1-NEXT: [[TMP39:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: store i32 [[TMP39]], ptr [[I_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[TMP40:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[IDXPROM_I:%.*]] = sext i32 [[TMP40]] to i64
+// CHECK1-NEXT: [[ARRAYIDX_I:%.*]] = getelementptr inbounds i16, ptr [[TMP35]], i64 [[IDXPROM_I]]
+// CHECK1-NEXT: [[TMP41:%.*]] = load i16, ptr [[ARRAYIDX_I]], align 2, !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[CONV3_I:%.*]] = sext i16 [[TMP41]] to i32
+// CHECK1-NEXT: [[TMP42:%.*]] = load i32, ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[ADD4_I:%.*]] = add nsw i32 [[TMP42]], [[CONV3_I]]
// CHECK1-NEXT: store i32 [[ADD4_I]], ptr [[TMP29]], align 4, !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
-// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP44]], 1
-// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META18]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[TMP43:%.*]] = load i32, ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
+// CHECK1-NEXT: [[ADD5_I:%.*]] = add nsw i32 [[TMP43]], 1
+// CHECK1-NEXT: store i32 [[ADD5_I]], ptr [[DOTOMP_IV_I]], align 4, !noalias [[META17]], !llvm.access.group [[ACC_GRP19]]
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND_I]], !llvm.loop [[LOOP20:![0-9]+]]
// CHECK1: .omp_outlined..exit:
// CHECK1-NEXT: ret i32 0
@@ -706,32 +705,32 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: store i32 [[CONV]], ptr [[DOTOMP_IV]], align 4
// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK3: omp.inner.for.cond:
-// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2:![0-9]+]]
+// CHECK3-NEXT: [[TMP4:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1:![0-9]+]]
// CHECK3-NEXT: [[CONV2:%.*]] = sext i32 [[TMP4]] to i64
-// CHECK3-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP5:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[CMP:%.*]] = icmp ule i64 [[CONV2]], [[TMP5]]
// CHECK3-NEXT: br i1 [[CMP]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK3: omp.inner.for.body:
-// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP6:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP6]], 1
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
-// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP2]]
-// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: store i32 [[ADD]], ptr [[I]], align 4, !llvm.access.group [[ACC_GRP1]]
+// CHECK3-NEXT: [[TMP7:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP7]] to i64
// CHECK3-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i16, ptr [[VLA]], i64 [[IDXPROM]]
-// CHECK3-NEXT: [[TMP8:%.*]] = load i16, ptr [[ARRAYIDX]], align 2, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP8:%.*]] = load i16, ptr [[ARRAYIDX]], align 2, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[CONV3:%.*]] = sext i16 [[TMP8]] to i32
-// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP9:%.*]] = load i32, ptr [[A]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP9]], [[CONV3]]
-// CHECK3-NEXT: store i32 [[ADD4]], ptr [[A]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: store i32 [[ADD4]], ptr [[A]], align 4, !llvm.access.group [[ACC_GRP1]]
// 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: [[TMP10:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
+// CHECK3-NEXT: [[TMP10:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
// CHECK3-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP10]], 1
-// CHECK3-NEXT: store i32 [[ADD5]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP2]]
-// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP3:![0-9]+]]
+// CHECK3-NEXT: store i32 [[ADD5]], ptr [[DOTOMP_IV]], align 4, !llvm.access.group [[ACC_GRP1]]
+// CHECK3-NEXT: br label [[OMP_INNER_FOR_COND]], !llvm.loop [[LOOP2:![0-9]+]]
// CHECK3: omp.inner.for.end:
// CHECK3-NEXT: store i32 5, ptr [[I]], align 4
// CHECK3-NEXT: store i32 0, ptr [[RETVAL]], align 4
diff --git a/clang/test/OpenMP/taskloop_simd_reduction_codegen.cpp b/clang/test/OpenMP/taskloop_simd_reduction_codegen.cpp
index 09ddedbbc7f3c..b3764d7c977dd 100644
--- a/clang/test/OpenMP/taskloop_simd_reduction_codegen.cpp
+++ b/clang/test/OpenMP/taskloop_simd_reduction_codegen.cpp
@@ -90,10 +90,9 @@ sum = 0.0;
// CHECK-DAG: [[TMP32:%.*]] = ptrtoaddr ptr [[ARRAYIDX6]] to i64
// CHECK-DAG: [[TMP33:%.*]] = ptrtoaddr ptr [[ARRAYIDX5]] to i64
// CHECK-DAG: [[TMP34:%.*]] = sub i64 [[TMP32]], [[TMP33]]
-// CHECK-DAG: [[TMP35:%.*]] = sdiv exact i64 [[TMP34]], 4
-// CHECK-DAG: [[TMP36:%.*]] = add nuw i64 [[TMP35]], 1
-// CHECK-DAG: [[TMP37:%.*]] = mul nuw i64 [[TMP36]], ptrtoint (ptr getelementptr (%struct.S, ptr null, i32 1) to i64)
-// CHECK-DAG: store i64 [[TMP37]], ptr [[TMP38:%[^,]+]],
+// CHECK-DAG: [[TMP35:%.*]] = add nuw i64 [[TMP34]], 4
+// CHECK-DAG: [[TMP36:%.*]] = udiv exact i64 [[TMP35]], 4
+// CHECK-DAG: store i64 [[TMP35]], ptr [[TMP38:%[^,]+]],
// CHECK-DAG: [[TMP38]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 2
// CHECK-DAG: [[TMP39:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_4]], i32 0, i32 3
// CHECK-DAG: store ptr @[[RED_INIT2:.+]], ptr [[TMP39]],
@@ -123,7 +122,7 @@ sum = 0.0;
// CHECK-DAG: store ptr [[VLA]], ptr [[TMP52:%[^,]+]],
// CHECK-DAG: [[TMP52]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 1
// CHECK-DAG: [[TMP54:%.*]] = mul nuw i64 [[TMP2]], 4
-// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], ptrtoint (ptr getelementptr (float, ptr null, i32 1) to i64)
+// CHECK-DAG: [[TMP55:%.*]] = udiv exact i64 [[TMP54]], 4
// CHECK-DAG: store i64 [[TMP54]], ptr [[TMP56:%[^,]+]],
// CHECK-DAG: [[TMP56]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 2
// CHECK-DAG: [[TMP57:%.*]] = getelementptr inbounds nuw %struct.kmp_taskred_input_t, ptr [[DOTRD_INPUT_GEP_8]], i32 0, i32 3
diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
index 3563df4893970..a53426bf068ea 100644
--- a/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
@@ -116,203 +116,201 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP8:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP9:%.*]] = sub i64 [[TMP7]], [[TMP8]]
// CHECK1-NEXT: [[TMP10:%.*]] = add nuw i64 [[TMP9]], 1
-// CHECK1-NEXT: [[TMP11:%.*]] = mul nuw i64 [[TMP10]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP12:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP12]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP11:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP11]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP10]], align 16
// CHECK1-NEXT: store i64 [[TMP10]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP13:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP13]]
+// CHECK1-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP13]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP12]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[TMP14]], align 8
-// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[TMP15]] to i64
-// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP18:%.*]] = sub i64 [[TMP16]], [[TMP17]]
-// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP18]]
+// CHECK1-NEXT: [[TMP13:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP14:%.*]] = load ptr, ptr [[TMP13]], align 8
+// CHECK1-NEXT: [[TMP15:%.*]] = ptrtoaddr ptr [[TMP14]] to i64
+// CHECK1-NEXT: [[TMP16:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP17:%.*]] = sub i64 [[TMP15]], [[TMP16]]
+// CHECK1-NEXT: [[TMP18:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP17]]
// CHECK1-NEXT: store ptr [[_TMP5]], ptr [[TMP]], align 8
-// CHECK1-NEXT: store ptr [[TMP19]], ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP18]], ptr [[_TMP5]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP20]], align 8
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP21]], align 8
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP26]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP19:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP19]], align 8
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP20]], align 8
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init., ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb., ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP25]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_6:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP28]], i64 0
-// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP29]], i64 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP31:%.*]] = sext i32 [[TMP30]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP31]]
-// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP32]], i64 9
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP33]], i64 [[LB_ADD_LEN9]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP34]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
-// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
-// CHECK1-NEXT: [[TMP37:%.*]] = sub i64 [[TMP35]], [[TMP36]]
-// CHECK1-NEXT: [[TMP38:%.*]] = add nuw i64 [[TMP37]], 1
-// CHECK1-NEXT: [[TMP39:%.*]] = mul nuw i64 [[TMP38]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP41]], align 8
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
-// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP27:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX7:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP27]], i64 0
+// CHECK1-NEXT: [[TMP28:%.*]] = load ptr, ptr [[ARRAYIDX7]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP28]], i64 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP30:%.*]] = sext i32 [[TMP29]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN9:%.*]] = add nsw i64 -1, [[TMP30]]
+// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX10:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP31]], i64 9
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[ARRAYIDX10]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP32]], i64 [[LB_ADD_LEN9]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX8]], ptr [[TMP33]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = ptrtoaddr ptr [[ARRAYIDX11]] to i64
+// CHECK1-NEXT: [[TMP35:%.*]] = ptrtoaddr ptr [[ARRAYIDX8]] to i64
+// CHECK1-NEXT: [[TMP36:%.*]] = sub i64 [[TMP34]], [[TMP35]]
+// CHECK1-NEXT: [[TMP37:%.*]] = add nuw i64 [[TMP36]], 1
+// CHECK1-NEXT: [[TMP38:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP37]], ptr [[TMP38]], align 8
+// CHECK1-NEXT: [[TMP39:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..1, ptr [[TMP39]], align 8
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..2, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T]], ptr [[DOTRD_INPUT_GEP_6]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
+// CHECK1-NEXT: [[TMP45:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP44]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP45]], ptr [[DOTTASK_RED_]], align 8
// CHECK1-NEXT: store i64 0, ptr [[DOTOMP_COMB_LB]], align 8
// CHECK1-NEXT: store i64 9, ptr [[DOTOMP_COMB_UB]], align 8
// CHECK1-NEXT: store i64 1, ptr [[DOTOMP_STRIDE]], align 8
// CHECK1-NEXT: store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
-// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP49]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
+// CHECK1-NEXT: [[TMP46:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP47:%.*]] = load i32, ptr [[TMP46]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB2:[0-9]+]], i32 [[TMP47]], i32 92, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP48:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP48]], 9
// 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: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP49]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP50]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP51]], [[TMP52]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
-// CHECK1-NEXT: [[TMP57:%.*]] = load ptr, ptr [[TMP]], align 8
-// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l16.omp_outlined.omp_outlined, i64 [[TMP55]], i64 [[TMP56]], ptr [[ARGC1]], ptr [[TMP57]])
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_COMB_LB]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_COMB_UB]], align 8
+// CHECK1-NEXT: [[TMP55:%.*]] = load ptr, ptr [[TMP]], align 8
+// CHECK1-NEXT: call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr @[[GLOB1]], i32 4, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l16.omp_outlined.omp_outlined, i64 [[TMP53]], i64 [[TMP54]], ptr [[ARGC1]], ptr [[TMP55]])
// CHECK1-NEXT: br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1: omp.inner.for.inc:
-// CHECK1-NEXT: [[TMP58:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = load i64, ptr [[DOTOMP_STRIDE]], align 8
-// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 [[TMP58]], [[TMP59]]
+// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = load i64, ptr [[DOTOMP_STRIDE]], align 8
+// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 [[TMP56]], [[TMP57]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP58:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP59:%.*]] = load i32, ptr [[TMP58]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP59]])
// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB2]], i32 [[TMP61]])
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP63]], i32 1)
-// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP64]], align 8
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP65]], align 8
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP67:%.*]] = inttoptr i64 [[TMP10]] to ptr
-// CHECK1-NEXT: store ptr [[TMP67]], ptr [[TMP66]], align 8
-// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP68]], align 4
-// CHECK1-NEXT: [[TMP70:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4:[0-9]+]], i32 [[TMP69]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l16.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP70]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP61]], i32 1)
+// CHECK1-NEXT: [[TMP62:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP62]], align 8
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP63]], align 8
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP65:%.*]] = inttoptr i64 [[TMP10]] to ptr
+// CHECK1-NEXT: store ptr [[TMP65]], ptr [[TMP64]], align 8
+// CHECK1-NEXT: [[TMP66:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP67:%.*]] = load i32, ptr [[TMP66]], align 4
+// CHECK1-NEXT: [[TMP68:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4:[0-9]+]], i32 [[TMP67]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l16.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP68]], 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: [[TMP71:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[TMP71]], [[TMP72]]
+// CHECK1-NEXT: [[TMP69:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[TMP69]], [[TMP70]]
// CHECK1-NEXT: store i32 [[ADD14]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP73:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP73]]
+// CHECK1-NEXT: [[TMP71:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP71]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE21:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST15:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT19:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP74:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP74]] to i32
-// CHECK1-NEXT: [[TMP75:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV16:%.*]] = sext i8 [[TMP75]] to i32
+// CHECK1-NEXT: [[TMP72:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP72]] to i32
+// CHECK1-NEXT: [[TMP73:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV16:%.*]] = sext i8 [[TMP73]] to i32
// CHECK1-NEXT: [[ADD17:%.*]] = add nsw i32 [[CONV]], [[CONV16]]
// CHECK1-NEXT: [[CONV18:%.*]] = trunc i32 [[ADD17]] to i8
// CHECK1-NEXT: store i8 [[CONV18]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT19]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST15]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT19]], [[TMP73]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE20:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT19]], [[TMP71]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE20]], label [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done21:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP69]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP67]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP77:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP76]] monotonic, align 4
-// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY22:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP78]]
+// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP75:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP74]] monotonic, align 4
+// CHECK1-NEXT: [[TMP76:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP10]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY22:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP76]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY22]], label [[OMP_ARRAYCPY_DONE35:%.*]], label [[OMP_ARRAYCPY_BODY23:%.*]]
// CHECK1: omp.arraycpy.body23:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST24:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT33:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST25:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT32:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
-// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP79]] to i32
+// CHECK1-NEXT: [[TMP77:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
+// CHECK1-NEXT: [[CONV26:%.*]] = sext i8 [[TMP77]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP80:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY23]] ], [ [[TMP85:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP80]], ptr [[_TMP27]], align 1
-// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[_TMP27]], align 1
-// CHECK1-NEXT: [[CONV28:%.*]] = sext i8 [[TMP81]] to i32
-// CHECK1-NEXT: [[TMP82:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP82]] to i32
+// CHECK1-NEXT: [[TMP78:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY23]] ], [ [[TMP83:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP78]], ptr [[_TMP27]], align 1
+// CHECK1-NEXT: [[TMP79:%.*]] = load i8, ptr [[_TMP27]], align 1
+// CHECK1-NEXT: [[CONV28:%.*]] = sext i8 [[TMP79]] to i32
+// CHECK1-NEXT: [[TMP80:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP80]] to i32
// CHECK1-NEXT: [[ADD30:%.*]] = add nsw i32 [[CONV28]], [[CONV29]]
// CHECK1-NEXT: [[CONV31:%.*]] = trunc i32 [[ADD30]] to i8
// CHECK1-NEXT: store i8 [[CONV31]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP83:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP84:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]], i8 [[TMP80]], i8 [[TMP83]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP85]] = extractvalue { i8, i1 } [[TMP84]], 0
-// CHECK1-NEXT: [[TMP86:%.*]] = extractvalue { i8, i1 } [[TMP84]], 1
-// CHECK1-NEXT: br i1 [[TMP86]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP81:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP82:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]], i8 [[TMP78]], i8 [[TMP81]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP83]] = extractvalue { i8, i1 } [[TMP82]], 0
+// CHECK1-NEXT: [[TMP84:%.*]] = extractvalue { i8, i1 } [[TMP82]], 1
+// CHECK1-NEXT: br i1 [[TMP84]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT32]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST25]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST24]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE34:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT32]], [[TMP78]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE34:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT32]], [[TMP76]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE34]], label [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_BODY23]]
// CHECK1: omp.arraycpy.done35:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP87:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP87]])
+// CHECK1-NEXT: [[TMP85:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP85]])
// CHECK1-NEXT: ret void
//
//
@@ -457,223 +455,221 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP10:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
// CHECK1-NEXT: [[TMP11:%.*]] = sub i64 [[TMP9]], [[TMP10]]
// CHECK1-NEXT: [[TMP12:%.*]] = add nuw i64 [[TMP11]], 1
-// CHECK1-NEXT: [[TMP13:%.*]] = mul nuw i64 [[TMP12]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP14:%.*]] = call ptr @llvm.stacksave.p0()
-// CHECK1-NEXT: store ptr [[TMP14]], ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: [[TMP13:%.*]] = call ptr @llvm.stacksave.p0()
+// CHECK1-NEXT: store ptr [[TMP13]], ptr [[SAVED_STACK]], align 8
// CHECK1-NEXT: [[VLA:%.*]] = alloca i8, i64 [[TMP12]], align 16
// CHECK1-NEXT: store i64 [[TMP12]], ptr [[__VLA_EXPR0]], align 8
-// CHECK1-NEXT: [[TMP15:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP15]]
+// CHECK1-NEXT: [[TMP14:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYINIT_ISEMPTY:%.*]] = icmp eq ptr [[VLA]], [[TMP14]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYINIT_ISEMPTY]], label [[OMP_ARRAYINIT_DONE:%.*]], label [[OMP_ARRAYINIT_BODY:%.*]]
// CHECK1: omp.arrayinit.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[ENTRY:%.*]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT:%.*]], [[OMP_ARRAYINIT_BODY]] ]
// CHECK1-NEXT: store i8 0, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP15]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT]], [[TMP14]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE]], label [[OMP_ARRAYINIT_DONE]], label [[OMP_ARRAYINIT_BODY]]
// CHECK1: omp.arrayinit.done:
-// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[TMP17:%.*]] = load ptr, ptr [[TMP16]], align 8
-// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[TMP17]] to i64
-// CHECK1-NEXT: [[TMP19:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
-// CHECK1-NEXT: [[TMP20:%.*]] = sub i64 [[TMP18]], [[TMP19]]
-// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP20]]
+// CHECK1-NEXT: [[TMP15:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[TMP16:%.*]] = load ptr, ptr [[TMP15]], align 8
+// CHECK1-NEXT: [[TMP17:%.*]] = ptrtoaddr ptr [[TMP16]] to i64
+// CHECK1-NEXT: [[TMP18:%.*]] = ptrtoaddr ptr [[ARRAYIDX2]] to i64
+// CHECK1-NEXT: [[TMP19:%.*]] = sub i64 [[TMP17]], [[TMP18]]
+// CHECK1-NEXT: [[TMP20:%.*]] = getelementptr i8, ptr [[VLA]], i64 [[TMP19]]
// CHECK1-NEXT: store ptr [[_TMP6]], ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP21]], ptr [[_TMP6]], align 8
+// CHECK1-NEXT: store ptr [[TMP20]], ptr [[_TMP6]], align 8
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T_0]]], ptr [[DOTRD_INPUT_]], i64 0, i64 0
-// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP22]], align 8
-// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP23]], align 8
-// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
-// CHECK1-NEXT: store i64 4, ptr [[TMP24]], align 8
-// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..3, ptr [[TMP25]], align 8
-// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP26]], align 8
-// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..4, ptr [[TMP27]], align 8
-// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
-// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP28]], i8 0, i64 4, i1 false)
+// CHECK1-NEXT: [[TMP21:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP21]], align 8
+// CHECK1-NEXT: [[TMP22:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[TMP0]], ptr [[TMP22]], align 8
+// CHECK1-NEXT: [[TMP23:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 2
+// CHECK1-NEXT: store i64 4, ptr [[TMP23]], align 8
+// CHECK1-NEXT: [[TMP24:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..3, ptr [[TMP24]], align 8
+// CHECK1-NEXT: [[TMP25:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP25]], align 8
+// CHECK1-NEXT: [[TMP26:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..4, ptr [[TMP26]], align 8
+// CHECK1-NEXT: [[TMP27:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_]], i32 0, i32 6
+// CHECK1-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[TMP27]], i8 0, i64 4, i1 false)
// CHECK1-NEXT: [[DOTRD_INPUT_GEP_7:%.*]] = getelementptr inbounds nuw [2 x [[STRUCT_KMP_TASKRED_INPUT_T_0]]], ptr [[DOTRD_INPUT_]], i64 0, i64 1
-// CHECK1-NEXT: [[TMP29:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP30]], i64 0
-// CHECK1-NEXT: [[TMP31:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
-// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP31]], i64 0
-// CHECK1-NEXT: [[TMP32:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP33:%.*]] = sext i32 [[TMP32]] to i64
-// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP33]]
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
-// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP34]], i64 9
-// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
-// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP35]], i64 [[LB_ADD_LEN10]]
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP29]], align 8
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
-// CHECK1-NEXT: [[TMP40:%.*]] = add nuw i64 [[TMP39]], 1
-// CHECK1-NEXT: [[TMP41:%.*]] = mul nuw i64 [[TMP40]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
-// CHECK1-NEXT: store i64 [[TMP41]], ptr [[TMP42]], align 8
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
-// CHECK1-NEXT: store ptr @.red_init..5, ptr [[TMP43]], align 8
-// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
-// CHECK1-NEXT: store ptr null, ptr [[TMP44]], align 8
-// CHECK1-NEXT: [[TMP45:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
-// CHECK1-NEXT: store ptr @.red_comb..6, ptr [[TMP45]], align 8
-// CHECK1-NEXT: [[TMP46:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
-// CHECK1-NEXT: store i32 1, ptr [[TMP46]], align 8
-// CHECK1-NEXT: [[TMP47:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP48:%.*]] = load i32, ptr [[TMP47]], align 4
-// CHECK1-NEXT: [[TMP49:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP48]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
-// CHECK1-NEXT: store ptr [[TMP49]], ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: [[TMP50:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP51:%.*]] = load i32, ptr [[TMP50]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP51]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
-// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP52]], 9
+// CHECK1-NEXT: [[TMP28:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP29:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX8:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP29]], i64 0
+// CHECK1-NEXT: [[TMP30:%.*]] = load ptr, ptr [[ARRAYIDX8]], align 8
+// CHECK1-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP30]], i64 0
+// CHECK1-NEXT: [[TMP31:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP32:%.*]] = sext i32 [[TMP31]] to i64
+// CHECK1-NEXT: [[LB_ADD_LEN10:%.*]] = add nsw i64 -1, [[TMP32]]
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[ARGV_ADDR]], align 8
+// CHECK1-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds nuw ptr, ptr [[TMP33]], i64 9
+// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[ARRAYIDX11]], align 8
+// CHECK1-NEXT: [[ARRAYIDX12:%.*]] = getelementptr inbounds nuw i8, ptr [[TMP34]], i64 [[LB_ADD_LEN10]]
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP28]], align 8
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARRAYIDX9]], ptr [[TMP35]], align 8
+// CHECK1-NEXT: [[TMP36:%.*]] = ptrtoaddr ptr [[ARRAYIDX12]] to i64
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[ARRAYIDX9]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = sub i64 [[TMP36]], [[TMP37]]
+// CHECK1-NEXT: [[TMP39:%.*]] = add nuw i64 [[TMP38]], 1
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 2
+// CHECK1-NEXT: store i64 [[TMP39]], ptr [[TMP40]], align 8
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 3
+// CHECK1-NEXT: store ptr @.red_init..5, ptr [[TMP41]], align 8
+// CHECK1-NEXT: [[TMP42:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 4
+// CHECK1-NEXT: store ptr null, ptr [[TMP42]], align 8
+// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 5
+// CHECK1-NEXT: store ptr @.red_comb..6, ptr [[TMP43]], align 8
+// CHECK1-NEXT: [[TMP44:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASKRED_INPUT_T_0]], ptr [[DOTRD_INPUT_GEP_7]], i32 0, i32 6
+// CHECK1-NEXT: store i32 1, ptr [[TMP44]], align 8
+// CHECK1-NEXT: [[TMP45:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP46:%.*]] = load i32, ptr [[TMP45]], align 4
+// CHECK1-NEXT: [[TMP47:%.*]] = call ptr @__kmpc_taskred_modifier_init(ptr @[[GLOB1]], i32 [[TMP46]], i32 1, i32 2, ptr [[DOTRD_INPUT_]])
+// CHECK1-NEXT: store ptr [[TMP47]], ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: [[TMP48:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP49:%.*]] = load i32, ptr [[TMP48]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_init_8(ptr @[[GLOB3:[0-9]+]], i32 [[TMP49]], i32 34, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i64 1, i64 1)
+// CHECK1-NEXT: [[TMP50:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP:%.*]] = icmp sgt i64 [[TMP50]], 9
// 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: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[TMP51:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
// CHECK1-NEXT: br label [[COND_END]]
// CHECK1: cond.end:
-// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP53]], [[COND_FALSE]] ]
+// CHECK1-NEXT: [[COND:%.*]] = phi i64 [ 9, [[COND_TRUE]] ], [ [[TMP51]], [[COND_FALSE]] ]
// CHECK1-NEXT: store i64 [[COND]], ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
-// CHECK1-NEXT: store i64 [[TMP54]], ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP52:%.*]] = load i64, ptr [[DOTOMP_LB]], align 8
+// CHECK1-NEXT: store i64 [[TMP52]], ptr [[DOTOMP_IV]], align 8
// CHECK1-NEXT: br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1: omp.inner.for.cond:
-// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[TMP56:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
-// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP55]], [[TMP56]]
+// CHECK1-NEXT: [[TMP53:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[TMP54:%.*]] = load i64, ptr [[DOTOMP_UB]], align 8
+// CHECK1-NEXT: [[CMP13:%.*]] = icmp sle i64 [[TMP53]], [[TMP54]]
// CHECK1-NEXT: br i1 [[CMP13]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_COND_CLEANUP:%.*]]
// CHECK1: omp.inner.for.cond.cleanup:
// CHECK1-NEXT: br label [[OMP_INNER_FOR_END:%.*]]
// CHECK1: omp.inner.for.body:
-// CHECK1-NEXT: [[TMP57:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP57]], 1
+// CHECK1-NEXT: [[TMP55:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i64 [[TMP55]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i64 0, [[MUL]]
// CHECK1-NEXT: store i64 [[ADD]], ptr [[I]], align 8
-// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
-// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP58]], align 8
-// CHECK1-NEXT: [[TMP59:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP59]], align 8
-// CHECK1-NEXT: [[TMP60:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
-// CHECK1-NEXT: [[TMP61:%.*]] = load ptr, ptr [[_TMP5]], align 8
-// CHECK1-NEXT: store ptr [[TMP61]], ptr [[TMP60]], align 8
-// CHECK1-NEXT: [[TMP62:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP63:%.*]] = load i32, ptr [[TMP62]], align 4
-// CHECK1-NEXT: [[TMP64:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP63]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
-// CHECK1-NEXT: [[TMP65:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP64]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP65]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP67:%.*]] = load ptr, ptr [[TMP66]], align 8
-// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP67]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
-// CHECK1-NEXT: [[TMP68:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP64]], i32 0, i32 1
-// CHECK1-NEXT: [[TMP69:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP68]], i32 0, i32 0
-// CHECK1-NEXT: [[TMP70:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
-// CHECK1-NEXT: store ptr [[TMP70]], ptr [[TMP69]], align 8
-// CHECK1-NEXT: [[TMP71:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP72:%.*]] = load i32, ptr [[TMP71]], align 4
-// CHECK1-NEXT: [[TMP73:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP72]], ptr [[TMP64]])
+// CHECK1-NEXT: [[TMP56:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 0
+// CHECK1-NEXT: store ptr [[DOTTASK_RED_]], ptr [[TMP56]], align 8
+// CHECK1-NEXT: [[TMP57:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 1
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP57]], align 8
+// CHECK1-NEXT: [[TMP58:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[AGG_CAPTURED]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP59:%.*]] = load ptr, ptr [[_TMP5]], align 8
+// CHECK1-NEXT: store ptr [[TMP59]], ptr [[TMP58]], align 8
+// CHECK1-NEXT: [[TMP60:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP61:%.*]] = load i32, ptr [[TMP60]], align 4
+// CHECK1-NEXT: [[TMP62:%.*]] = call ptr @__kmpc_omp_task_alloc(ptr @[[GLOB1]], i32 [[TMP61]], i32 1, i64 48, i64 24, ptr @.omp_task_entry.)
+// CHECK1-NEXT: [[TMP63:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES:%.*]], ptr [[TMP62]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP64:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T:%.*]], ptr [[TMP63]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP65:%.*]] = load ptr, ptr [[TMP64]], align 8
+// CHECK1-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[TMP65]], ptr align 8 [[AGG_CAPTURED]], i64 24, i1 false)
+// CHECK1-NEXT: [[TMP66:%.*]] = getelementptr inbounds nuw [[STRUCT_KMP_TASK_T_WITH_PRIVATES]], ptr [[TMP62]], i32 0, i32 1
+// CHECK1-NEXT: [[TMP67:%.*]] = getelementptr inbounds nuw [[STRUCT__KMP_PRIVATES_T:%.*]], ptr [[TMP66]], i32 0, i32 0
+// CHECK1-NEXT: [[TMP68:%.*]] = load ptr, ptr [[DOTTASK_RED_]], align 8
+// CHECK1-NEXT: store ptr [[TMP68]], ptr [[TMP67]], align 8
+// CHECK1-NEXT: [[TMP69:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP70:%.*]] = load i32, ptr [[TMP69]], align 4
+// CHECK1-NEXT: [[TMP71:%.*]] = call i32 @__kmpc_omp_task(ptr @[[GLOB1]], i32 [[TMP70]], ptr [[TMP62]])
// 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: [[TMP74:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
-// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP74]], 1
+// CHECK1-NEXT: [[TMP72:%.*]] = load i64, ptr [[DOTOMP_IV]], align 8
+// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i64 [[TMP72]], 1
// CHECK1-NEXT: store i64 [[ADD14]], ptr [[DOTOMP_IV]], align 8
// 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: [[TMP73:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP74:%.*]] = load i32, ptr [[TMP73]], align 4
+// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP74]])
// CHECK1-NEXT: [[TMP75:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT: [[TMP76:%.*]] = load i32, ptr [[TMP75]], align 4
-// CHECK1-NEXT: call void @__kmpc_for_static_fini(ptr @[[GLOB3]], i32 [[TMP76]])
-// CHECK1-NEXT: [[TMP77:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP78:%.*]] = load i32, ptr [[TMP77]], align 4
-// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP78]], i32 1)
-// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
-// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP79]], align 8
-// CHECK1-NEXT: [[TMP80:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
-// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP80]], align 8
-// CHECK1-NEXT: [[TMP81:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
-// CHECK1-NEXT: [[TMP82:%.*]] = inttoptr i64 [[TMP12]] to ptr
-// CHECK1-NEXT: store ptr [[TMP82]], ptr [[TMP81]], align 8
-// CHECK1-NEXT: [[TMP83:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
-// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP83]], align 4
-// CHECK1-NEXT: [[TMP85:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP84]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l16.omp_outlined.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
-// CHECK1-NEXT: switch i32 [[TMP85]], label [[DOTOMP_REDUCTION_DEFAULT:%.*]] [
+// CHECK1-NEXT: call void @__kmpc_task_reduction_modifier_fini(ptr @[[GLOB1]], i32 [[TMP76]], i32 1)
+// CHECK1-NEXT: [[TMP77:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 0
+// CHECK1-NEXT: store ptr [[ARGC1]], ptr [[TMP77]], align 8
+// CHECK1-NEXT: [[TMP78:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 1
+// CHECK1-NEXT: store ptr [[VLA]], ptr [[TMP78]], align 8
+// CHECK1-NEXT: [[TMP79:%.*]] = getelementptr inbounds [3 x ptr], ptr [[DOTOMP_REDUCTION_RED_LIST]], i64 0, i64 2
+// CHECK1-NEXT: [[TMP80:%.*]] = inttoptr i64 [[TMP12]] to ptr
+// CHECK1-NEXT: store ptr [[TMP80]], ptr [[TMP79]], align 8
+// CHECK1-NEXT: [[TMP81:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
+// CHECK1-NEXT: [[TMP82:%.*]] = load i32, ptr [[TMP81]], align 4
+// CHECK1-NEXT: [[TMP83:%.*]] = call i32 @__kmpc_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP82]], i32 2, i64 24, ptr [[DOTOMP_REDUCTION_RED_LIST]], ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l16.omp_outlined.omp_outlined.omp.reduction.reduction_func, ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: switch i32 [[TMP83]], 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: [[TMP86:%.*]] = load i32, ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP87:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP86]], [[TMP87]]
+// CHECK1-NEXT: [[TMP84:%.*]] = load i32, ptr [[TMP0]], align 4
+// CHECK1-NEXT: [[TMP85:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[ADD15:%.*]] = add nsw i32 [[TMP84]], [[TMP85]]
// CHECK1-NEXT: store i32 [[ADD15]], ptr [[TMP0]], align 4
-// CHECK1-NEXT: [[TMP88:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP88]]
+// CHECK1-NEXT: [[TMP86:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY]], label [[OMP_ARRAYCPY_DONE22:%.*]], label [[OMP_ARRAYCPY_BODY:%.*]]
// CHECK1: omp.arraycpy.body:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT:%.*]], [[OMP_ARRAYCPY_BODY]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST16:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE1]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT20:%.*]], [[OMP_ARRAYCPY_BODY]] ]
-// CHECK1-NEXT: [[TMP89:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
-// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP89]] to i32
-// CHECK1-NEXT: [[TMP90:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
-// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP90]] to i32
+// CHECK1-NEXT: [[TMP87:%.*]] = load i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
+// CHECK1-NEXT: [[CONV:%.*]] = sext i8 [[TMP87]] to i32
+// CHECK1-NEXT: [[TMP88:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], align 1
+// CHECK1-NEXT: [[CONV17:%.*]] = sext i8 [[TMP88]] to i32
// CHECK1-NEXT: [[ADD18:%.*]] = add nsw i32 [[CONV]], [[CONV17]]
// CHECK1-NEXT: [[CONV19:%.*]] = trunc i32 [[ADD18]] to i8
// CHECK1-NEXT: store i8 [[CONV19]], ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], align 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT20]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST16]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP88]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE21:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT20]], [[TMP86]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE21]], label [[OMP_ARRAYCPY_DONE22]], label [[OMP_ARRAYCPY_BODY]]
// CHECK1: omp.arraycpy.done22:
-// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP84]], ptr @.gomp_critical_user_.reduction.var)
+// CHECK1-NEXT: call void @__kmpc_end_reduce_nowait(ptr @[[GLOB4]], i32 [[TMP82]], ptr @.gomp_critical_user_.reduction.var)
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.case2:
-// CHECK1-NEXT: [[TMP91:%.*]] = load i32, ptr [[ARGC1]], align 4
-// CHECK1-NEXT: [[TMP92:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP91]] monotonic, align 4
-// CHECK1-NEXT: [[TMP93:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
-// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP93]]
+// CHECK1-NEXT: [[TMP89:%.*]] = load i32, ptr [[ARGC1]], align 4
+// CHECK1-NEXT: [[TMP90:%.*]] = atomicrmw add ptr [[TMP0]], i32 [[TMP89]] monotonic, align 4
+// CHECK1-NEXT: [[TMP91:%.*]] = getelementptr i8, ptr [[ARRAYIDX2]], i64 [[TMP12]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_ISEMPTY23:%.*]] = icmp eq ptr [[ARRAYIDX2]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_ISEMPTY23]], label [[OMP_ARRAYCPY_DONE36:%.*]], label [[OMP_ARRAYCPY_BODY24:%.*]]
// CHECK1: omp.arraycpy.body24:
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRCELEMENTPAST25:%.*]] = phi ptr [ [[VLA]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_SRC_ELEMENT34:%.*]], [[ATOMIC_EXIT:%.*]] ]
// CHECK1-NEXT: [[OMP_ARRAYCPY_DESTELEMENTPAST26:%.*]] = phi ptr [ [[ARRAYIDX2]], [[DOTOMP_REDUCTION_CASE2]] ], [ [[OMP_ARRAYCPY_DEST_ELEMENT33:%.*]], [[ATOMIC_EXIT]] ]
-// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP92:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV27:%.*]] = sext i8 [[TMP92]] to i32
// CHECK1-NEXT: [[ATOMIC_LOAD:%.*]] = load atomic i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]] monotonic, align 1
// CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]]
// CHECK1: atomic_cont:
-// CHECK1-NEXT: [[TMP95:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP100:%.*]], [[ATOMIC_CONT]] ]
-// CHECK1-NEXT: store i8 [[TMP95]], ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[_TMP28]], align 1
-// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP96]] to i32
-// CHECK1-NEXT: [[TMP97:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
-// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP97]] to i32
+// CHECK1-NEXT: [[TMP93:%.*]] = phi i8 [ [[ATOMIC_LOAD]], [[OMP_ARRAYCPY_BODY24]] ], [ [[TMP98:%.*]], [[ATOMIC_CONT]] ]
+// CHECK1-NEXT: store i8 [[TMP93]], ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[TMP94:%.*]] = load i8, ptr [[_TMP28]], align 1
+// CHECK1-NEXT: [[CONV29:%.*]] = sext i8 [[TMP94]] to i32
+// CHECK1-NEXT: [[TMP95:%.*]] = load i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], align 1
+// CHECK1-NEXT: [[CONV30:%.*]] = sext i8 [[TMP95]] to i32
// CHECK1-NEXT: [[ADD31:%.*]] = add nsw i32 [[CONV29]], [[CONV30]]
// CHECK1-NEXT: [[CONV32:%.*]] = trunc i32 [[ADD31]] to i8
// CHECK1-NEXT: store i8 [[CONV32]], ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP98:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
-// CHECK1-NEXT: [[TMP99:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP95]], i8 [[TMP98]] monotonic monotonic, align 1
-// CHECK1-NEXT: [[TMP100]] = extractvalue { i8, i1 } [[TMP99]], 0
-// CHECK1-NEXT: [[TMP101:%.*]] = extractvalue { i8, i1 } [[TMP99]], 1
-// CHECK1-NEXT: br i1 [[TMP101]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
+// CHECK1-NEXT: [[TMP96:%.*]] = load i8, ptr [[ATOMIC_TEMP]], align 1
+// CHECK1-NEXT: [[TMP97:%.*]] = cmpxchg ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i8 [[TMP93]], i8 [[TMP96]] monotonic monotonic, align 1
+// CHECK1-NEXT: [[TMP98]] = extractvalue { i8, i1 } [[TMP97]], 0
+// CHECK1-NEXT: [[TMP99:%.*]] = extractvalue { i8, i1 } [[TMP97]], 1
+// CHECK1-NEXT: br i1 [[TMP99]], label [[ATOMIC_EXIT]], label [[ATOMIC_CONT]]
// CHECK1: atomic_exit:
// CHECK1-NEXT: [[OMP_ARRAYCPY_DEST_ELEMENT33]] = getelementptr i8, ptr [[OMP_ARRAYCPY_DESTELEMENTPAST26]], i32 1
// CHECK1-NEXT: [[OMP_ARRAYCPY_SRC_ELEMENT34]] = getelementptr i8, ptr [[OMP_ARRAYCPY_SRCELEMENTPAST25]], i32 1
-// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP93]]
+// CHECK1-NEXT: [[OMP_ARRAYCPY_DONE35:%.*]] = icmp eq ptr [[OMP_ARRAYCPY_DEST_ELEMENT33]], [[TMP91]]
// CHECK1-NEXT: br i1 [[OMP_ARRAYCPY_DONE35]], label [[OMP_ARRAYCPY_DONE36]], label [[OMP_ARRAYCPY_BODY24]]
// CHECK1: omp.arraycpy.done36:
// CHECK1-NEXT: br label [[DOTOMP_REDUCTION_DEFAULT]]
// CHECK1: .omp.reduction.default:
-// CHECK1-NEXT: [[TMP102:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
-// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP102]])
+// CHECK1-NEXT: [[TMP100:%.*]] = load ptr, ptr [[SAVED_STACK]], align 8
+// CHECK1-NEXT: call void @llvm.stackrestore.p0(ptr [[TMP100]])
// CHECK1-NEXT: ret void
//
//
@@ -834,19 +830,18 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
-// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias [[META13]]
-// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP32:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP33:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP32]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP34:%.*]] = getelementptr inbounds nuw [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP35:%.*]] = load ptr, ptr [[TMP34]], align 8
// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
-// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP37]] to i64
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
-// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
+// CHECK1-NEXT: [[TMP37:%.*]] = ptrtoaddr ptr [[TMP36]] to i64
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoaddr ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = sub i64 [[TMP37]], [[TMP38]]
+// CHECK1-NEXT: [[TMP40:%.*]] = getelementptr i8, ptr [[TMP33]], i64 [[TMP39]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias [[META13]]
-// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
+// CHECK1-NEXT: store ptr [[TMP40]], ptr [[TMP4_I]], align 8, !noalias [[META13]]
// CHECK1-NEXT: ret i32 0
//
//
More information about the cfe-commits
mailing list