Vendor import of clang trunk r256633:vendor/clang/clang-trunk-r256633

https://llvm.org/svn/llvm-project/cfe/trunk@256633

1 files changed, 94 insertions, 83 deletions

diff --git a/test/OpenMP/parallel_reduction_codegen.cpp b/test/OpenMP/parallel_reduction_codegen.cpp
index 04d19ebea8c5..b9744b634118 100644
--- a/test/OpenMP/parallel_reduction_codegen.cpp
+++ b/test/OpenMP/parallel_reduction_codegen.cpp
@@ -8,7 +8,7 @@
 #ifndef HEADER
 #define HEADER
 
-volatile int g = 1212;
+volatile int g __attribute__((aligned(128))) = 1212;
 
 template <class T>
 struct S {
@@ -22,9 +22,6 @@ struct S {
 
 // CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
 // CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
-// CHECK-DAG: [[CAP_MAIN_TY:%.+]] = type { [2 x i{{[0-9]+}}]*, float*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float* }
-// CHECK-DAG: [[CAP_TMAIN_TY:%.+]] = type { [2 x i{{[0-9]+}}]*, i{{[0-9]+}}*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*, [[S_INT_TY]]*, i{{[0-9]+}}* }
-// CHECK-DAG: [[IMPLICIT_BARRIER_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 66, i32 0, i32 0, i8*
 // CHECK-DAG: [[REDUCTION_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 18, i32 0, i32 0, i8*
 // CHECK-DAG: [[REDUCTION_LOCK:@.+]] = common global [8 x i32] zeroinitializer
 
@@ -32,10 +29,10 @@ template <typename T>
 T tmain() {
   T t;
   S<T> test;
-  T t_var = T(), t_var1;
+  T t_var __attribute__((aligned(128))) = T(), t_var1 __attribute__((aligned(128)));
   T vec[] = {1, 2};
-  S<T> s_arr[] = {1, 2};
-  S<T> var(3), var1;
+  S<T> s_arr[]  = {1, 2};
+  S<T> var __attribute__((aligned(128))) (3), var1 __attribute__((aligned(128)));
 #pragma omp parallel reduction(+:t_var) reduction(&:var) reduction(&& : var1) reduction(min: t_var1)
   {
     vec[0] = t_var;
@@ -51,30 +48,24 @@ int main() {
   // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
   [&]() {
   // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
-  // LAMBDA: [[G_LOCAL_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[AGG_CAPTURED:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
-  // LAMBDA: store i{{[0-9]+}}* [[G]], i{{[0-9]+}}** [[G_LOCAL_REF]]
-  // LAMBDA: [[ARG:%.+]] = bitcast %{{.+}}* [[AGG_CAPTURED]] to i8*
-  // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 1, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}, i8* [[ARG]])
+  // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 1, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}, i32* [[G]])
 #pragma omp parallel reduction(+:g)
   {
-    // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* %{{.+}}, i32* %{{.+}}, %{{.+}}* [[ARG:%.+]])
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}})
     // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
 
     // Reduction list for runtime.
     // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
-    // LAMBDA: store %{{.+}}* [[ARG]], %{{.+}}** [[ARG_REF:%.+]],
-    // LAMBDA: [[ARG:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_REF]]
-    // LAMBDA: [[G_REF_ADDR:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
-    // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_REF_ADDR]]
-    // LAMBDA: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
+    // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_REF_ADDR:%.+]]
+    // LAMBDA: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[G_PRIVATE_ADDR]], align 128
     g = 1;
-    // LAMBDA: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
+    // LAMBDA: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]], align 128
     // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
     // LAMBDA: store i{{[0-9]+}}* [[G_PRIVATE_ADDR]], i{{[0-9]+}}** [[G_PRIVATE_ADDR_REF]]
     // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
 
-    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // LAMBDA: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i64 0, i64 0
     // LAMBDA: [[BITCAST:%.+]] = bitcast i32* [[G_PRIVATE_ADDR]] to i8*
     // LAMBDA: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // LAMBDA: call i32 @__kmpc_reduce_nowait(
@@ -112,31 +103,25 @@ int main() {
   // BLOCKS: call void {{%.+}}(i8
   ^{
   // BLOCKS: define{{.*}} internal{{.*}} void {{.+}}(i8*
-  // BLOCKS: [[G_LOCAL_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[AGG_CAPTURED:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
-  // BLOCKS: store i{{[0-9]+}}* [[G]], i{{[0-9]+}}** [[G_LOCAL_REF]]
-  // BLOCKS: [[ARG:%.+]] = bitcast %{{.+}}* [[AGG_CAPTURED]] to i8*
-  // BLOCKS: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 1, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}, i8* [[ARG]])
+  // BLOCKS: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 1, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}, i32* [[G]])
 #pragma omp parallel reduction(-:g)
   {
-    // BLOCKS: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* %{{.+}}, i32* %{{.+}}, %{{.+}}* [[ARG:%.+]])
+    // BLOCKS: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}})
     // BLOCKS: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
 
     // Reduction list for runtime.
     // BLOCKS: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
-    // BLOCKS: store %{{.+}}* [[ARG]], %{{.+}}** [[ARG_REF:%.+]],
-    // BLOCKS: [[ARG:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_REF]]
-    // BLOCKS: [[G_REF_ADDR:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
-    // BLOCKS: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_REF_ADDR]]
-    // BLOCKS: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
+    // BLOCKS: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_REF_ADDR:%.+]]
+    // BLOCKS: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[G_PRIVATE_ADDR]], align 128
     g = 1;
-    // BLOCKS: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
+    // BLOCKS: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]], align 128
     // BLOCKS-NOT: [[G]]{{[[^:word:]]}}
     // BLOCKS: i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
     // BLOCKS-NOT: [[G]]{{[[^:word:]]}}
     // BLOCKS: call void {{%.+}}(i8
 
-    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i32 0, i32 0
+    // BLOCKS: [[G_PRIV_REF:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[RED_LIST]], i64 0, i64 0
     // BLOCKS: [[BITCAST:%.+]] = bitcast i32* [[G_PRIVATE_ADDR]] to i8*
     // BLOCKS: store i8* [[BITCAST]], i8** [[G_PRIV_REF]],
     // BLOCKS: call i32 @__kmpc_reduce_nowait(
@@ -178,6 +163,12 @@ int main() {
     vec[0] = t_var;
     s_arr[0] = var;
   }
+  if (var1)
+#pragma omp parallel reduction(+ : t_var) reduction(& : var) reduction(&& : var1) reduction(min : t_var1)
+    while (1) {
+      vec[0] = t_var;
+      s_arr[0] = var;
+    }
   return tmain<int>();
 #endif
 }
@@ -185,13 +176,13 @@ int main() {
 // CHECK: define {{.*}}i{{[0-9]+}} @main()
 // CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
-// CHECK: %{{.+}} = bitcast [[CAP_MAIN_TY]]*
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[CAP_MAIN_TY]]*)* [[MAIN_MICROTASK:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i32]*, float*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float*)* [[MAIN_MICROTASK:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i32]*, float*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float*)* [[MAIN_MICROTASK1:@.+]] to void
 // CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT:@.+]]()
 // CHECK: call {{.*}} [[S_FLOAT_TY_DESTR:@.+]]([[S_FLOAT_TY]]*
 // CHECK: ret
 //
-// CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* [[GTID_ADDR:%.+]], i{{[0-9]+}}* %{{.+}}, [[CAP_MAIN_TY]]* %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca float,
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
@@ -202,23 +193,20 @@ int main() {
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK: [[T_VAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_MAIN_TY]], [[CAP_MAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[T_VAR_REF:%.+]] = load float*, float** [[T_VAR_PTR_REF]],
+// CHECK: [[T_VAR_REF:%.+]] = load float*, float** %
+// CHECK: [[VAR_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** %
+// CHECK: [[VAR1_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** %
+// CHECK: [[T_VAR1_REF:%.+]] = load float*, float** %
+
 // For + reduction operation initial value of private variable is 0.
 // CHECK: store float 0.0{{.+}}, float* [[T_VAR_PRIV]],
 
-// CHECK: [[VAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_MAIN_TY]], [[CAP_MAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[VAR_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PTR_REF:%.+]],
 // For & reduction operation initial value of private variable is ones in all bits.
 // CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
 
-// CHECK: [[VAR1_PTR_REF:%.+]] = getelementptr inbounds [[CAP_MAIN_TY]], [[CAP_MAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[VAR1_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PTR_REF:%.+]],
 // For && reduction operation initial value of private variable is 1.0.
 // CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[VAR1_PRIV]])
 
-// CHECK: [[T_VAR1_PTR_REF:%.+]] = getelementptr inbounds [[CAP_MAIN_TY]], [[CAP_MAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[T_VAR1_REF:%.+]] = load float*, float** [[T_VAR1_PTR_REF]],
 // For min reduction operation initial value of private variable is largest repesentable value.
 // CHECK: store float 0x47EFFFFFE0000000, float* [[T_VAR1_PRIV]],
 
@@ -226,16 +214,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast float* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast float* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -359,7 +347,6 @@ int main() {
 // break;
 // CHECK: br label %[[RED_DONE]]
 // CHECK: [[RED_DONE]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
 
 // CHECK-DAG: call {{.*}} [[S_FLOAT_TY_DESTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
 // CHECK-DAG: call {{.*}} [[S_FLOAT_TY_DESTR]]([[S_FLOAT_TY]]*
@@ -373,38 +360,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (float*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to float*
 // t_var_rhs = (float*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to float*
 
 // var_lhs = (S<float>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_FLOAT_TY]]*
 // var_rhs = (S<float>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_FLOAT_TY]]*
 
 // var1_lhs = (S<float>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_FLOAT_TY]]*
 // var1_rhs = (S<float>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_FLOAT_TY]]*
 
 // t_var1_lhs = (float*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to float*
 // t_var1_rhs = (float*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to float*
 
@@ -445,42 +432,67 @@ int main() {
 // CHECK: store float [[UP]], float* [[T_VAR1_LHS]],
 // CHECK: ret void
 
+// CHECK: define internal void [[MAIN_MICROTASK1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca float,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[T_VAR1_PRIV:%.+]] = alloca float,
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+
+// CHECK: [[T_VAR_REF:%.+]] = load float*, float** %
+// CHECK: [[VAR_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** %
+// CHECK: [[VAR1_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** %
+// CHECK: [[T_VAR1_REF:%.+]] = load float*, float** %
+
+// For + reduction operation initial value of private variable is 0.
+// CHECK: store float 0.0{{.+}}, float* [[T_VAR_PRIV]],
+
+// For & reduction operation initial value of private variable is ones in all bits.
+// CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+
+// For && reduction operation initial value of private variable is 1.0.
+// CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[VAR1_PRIV]])
+
+// For min reduction operation initial value of private variable is largest repesentable value.
+// CHECK: store float 0x47EFFFFFE0000000, float* [[T_VAR1_PRIV]],
+
+// CHECK-NOT: call i32 @__kmpc_reduce
+
+// CHECK: ret void
+
 // CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]()
 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
 // CHECK: call {{.*}} [[S_INT_TY_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
-// CHECK: %{{.+}} = bitcast [[CAP_TMAIN_TY]]*
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[CAP_TMAIN_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i32]*, i32*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*, [[S_INT_TY]]*, i32*)* [[TMAIN_MICROTASK:@.+]] to void
 // CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: ret
 //
-// CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* [[GTID_ADDR:%.+]], i{{[0-9]+}}* %{{.+}}, [[CAP_TMAIN_TY]]* %{{.+}})
-// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
-// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
-// CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_INT_TY]],
-// CHECK: [[T_VAR1_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}}, align 128
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]], align 128
+// CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_INT_TY]], align 128
+// CHECK: [[T_VAR1_PRIV:%.+]] = alloca i{{[0-9]+}}, align 128
 
 // Reduction list for runtime.
 // CHECK: [[RED_LIST:%.+]] = alloca [4 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK: [[T_VAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_TMAIN_TY]], [[CAP_TMAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_PTR_REF]],
+// CHECK: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %
+// CHECK: [[VAR_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** %
+// CHECK: [[VAR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** %
+// CHECK: [[T_VAR1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %
+
 // For + reduction operation initial value of private variable is 0.
 // CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[T_VAR_PRIV]],
 
-// CHECK: [[VAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_TMAIN_TY]], [[CAP_TMAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[VAR_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_PTR_REF:%.+]],
 // For & reduction operation initial value of private variable is ones in all bits.
 // CHECK: call {{.*}} [[S_INT_TY_CONSTR:@.+]]([[S_INT_TY]]* [[VAR_PRIV]])
 
-// CHECK: [[VAR1_PTR_REF:%.+]] = getelementptr inbounds [[CAP_TMAIN_TY]], [[CAP_TMAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[VAR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_PTR_REF:%.+]],
 // For && reduction operation initial value of private variable is 1.0.
 // CHECK: call {{.*}} [[S_INT_TY_CONSTR:@.+]]([[S_INT_TY]]* [[VAR1_PRIV]])
 
-// CHECK: [[T_VAR1_PTR_REF:%.+]] = getelementptr inbounds [[CAP_TMAIN_TY]], [[CAP_TMAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} {{[0-9]+}}
-// CHECK: [[T_VAR1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR1_PTR_REF]],
 // For min reduction operation initial value of private variable is largest repesentable value.
 // CHECK: store i{{[0-9]+}} 2147483647, i{{[0-9]+}}* [[T_VAR1_PRIV]],
 
@@ -488,16 +500,16 @@ int main() {
 
 // void *RedList[<n>] = {<ReductionVars>[0], ..., <ReductionVars>[<n>-1]};
 
-// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 0
+// CHECK: [[T_VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 0
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR_PRIV_REF]],
-// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 1
+// CHECK: [[VAR_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 1
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR_PRIV_REF]],
-// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 2
+// CHECK: [[VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 2
 // CHECK: [[BITCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[VAR1_PRIV_REF]],
-// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i32 0, i32 3
+// CHECK: [[T_VAR1_PRIV_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST]], i64 0, i64 3
 // CHECK: [[BITCAST:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR1_PRIV]] to i8*
 // CHECK: store i8* [[BITCAST]], i8** [[T_VAR1_PRIV_REF]],
 
@@ -594,7 +606,6 @@ int main() {
 // break;
 // CHECK: br label %[[RED_DONE]]
 // CHECK: [[RED_DONE]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
 
 // CHECK-DAG: call {{.*}} [[S_INT_TY_DESTR]]([[S_INT_TY]]* [[VAR_PRIV]])
 // CHECK-DAG: call {{.*}} [[S_INT_TY_DESTR]]([[S_INT_TY]]*
@@ -608,38 +619,38 @@ int main() {
 // }
 // CHECK: define internal void [[REDUCTION_FUNC]](i8*, i8*)
 // t_var_lhs = (i{{[0-9]+}}*)lhs[0];
-// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR_RHS_REF]],
 // CHECK: [[T_VAR_RHS:%.+]] = bitcast i8* [[T_VAR_RHS_VOID]] to i{{[0-9]+}}*
 // t_var_rhs = (i{{[0-9]+}}*)rhs[0];
-// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i32 0, i32 0
+// CHECK: [[T_VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS:%.+]], i64 0, i64 0
 // CHECK: [[T_VAR_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR_LHS_REF]],
 // CHECK: [[T_VAR_LHS:%.+]] = bitcast i8* [[T_VAR_LHS_VOID]] to i{{[0-9]+}}*
 
 // var_lhs = (S<i{{[0-9]+}}>*)lhs[1];
-// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 1
+// CHECK: [[VAR_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 1
 // CHECK: [[VAR_RHS_VOID:%.+]] = load i8*, i8** [[VAR_RHS_REF]],
 // CHECK: [[VAR_RHS:%.+]] = bitcast i8* [[VAR_RHS_VOID]] to [[S_INT_TY]]*
 // var_rhs = (S<i{{[0-9]+}}>*)rhs[1];
-// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 1
+// CHECK: [[VAR_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 1
 // CHECK: [[VAR_LHS_VOID:%.+]] = load i8*, i8** [[VAR_LHS_REF]],
 // CHECK: [[VAR_LHS:%.+]] = bitcast i8* [[VAR_LHS_VOID]] to [[S_INT_TY]]*
 
 // var1_lhs = (S<i{{[0-9]+}}>*)lhs[2];
-// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 2
+// CHECK: [[VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 2
 // CHECK: [[VAR1_RHS_VOID:%.+]] = load i8*, i8** [[VAR1_RHS_REF]],
 // CHECK: [[VAR1_RHS:%.+]] = bitcast i8* [[VAR1_RHS_VOID]] to [[S_INT_TY]]*
 // var1_rhs = (S<i{{[0-9]+}}>*)rhs[2];
-// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 2
+// CHECK: [[VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 2
 // CHECK: [[VAR1_LHS_VOID:%.+]] = load i8*, i8** [[VAR1_LHS_REF]],
 // CHECK: [[VAR1_LHS:%.+]] = bitcast i8* [[VAR1_LHS_VOID]] to [[S_INT_TY]]*
 
 // t_var1_lhs = (i{{[0-9]+}}*)lhs[3];
-// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_RHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_RHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_RHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_RHS_REF]],
 // CHECK: [[T_VAR1_RHS:%.+]] = bitcast i8* [[T_VAR1_RHS_VOID]] to i{{[0-9]+}}*
 // t_var1_rhs = (i{{[0-9]+}}*)rhs[3];
-// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i32 0, i32 3
+// CHECK: [[T_VAR1_LHS_REF:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[RED_LIST_LHS]], i64 0, i64 3
 // CHECK: [[T_VAR1_LHS_VOID:%.+]] = load i8*, i8** [[T_VAR1_LHS_REF]],
 // CHECK: [[T_VAR1_LHS:%.+]] = bitcast i8* [[T_VAR1_LHS_VOID]] to i{{[0-9]+}}*