Index: clang/lib/Sema/Sema.cpp
===================================================================
--- clang/lib/Sema/Sema.cpp
+++ clang/lib/Sema/Sema.cpp
@@ -923,10 +923,7 @@
                                  Pending.begin(), Pending.end());
   }
 
-  {
-    llvm::TimeTraceScope TimeScope("PerformPendingInstantiations");
-    PerformPendingInstantiations();
-  }
+  PerformPendingInstantiations();
 
   // Finalize analysis of OpenMP-specific constructs.
   if (LangOpts.OpenMP)
@@ -983,6 +980,8 @@
                                  LateParsedInstantiations.begin(),
                                  LateParsedInstantiations.end());
     LateParsedInstantiations.clear();
+
+    PerformPendingInstantiations();
   }
 
   DiagnoseUnterminatedPragmaPack();
Index: clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
===================================================================
--- clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
+++ clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
@@ -5854,6 +5854,8 @@
 /// Performs template instantiation for all implicit template
 /// instantiations we have seen until this point.
 void Sema::PerformPendingInstantiations(bool LocalOnly) {
+  llvm::TimeTraceScope TimeScope("PerformPendingInstantiations",
+                                 StringRef(""));
   while (!PendingLocalImplicitInstantiations.empty() ||
          (!LocalOnly && !PendingInstantiations.empty())) {
     PendingImplicitInstantiation Inst;
Index: clang/test/CodeGenCXX/vla-lambda-capturing.cpp
===================================================================
--- clang/test/CodeGenCXX/vla-lambda-capturing.cpp
+++ clang/test/CodeGenCXX/vla-lambda-capturing.cpp
@@ -1,6 +1,6 @@
-// RUN: %clang_cc1 %s -std=c++11 -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 %s -std=c++11 -emit-llvm -o - | FileCheck --check-prefixes=CHECK,NOPCH %s
 // RUN: %clang_cc1 %s -std=c++11 -emit-pch -o %t
-// RUN: %clang_cc1 %s -std=c++11 -include-pch %t -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 %s -std=c++11 -include-pch %t -emit-llvm -o - | FileCheck --check-prefixes=CHECK,PCH %s
 
 #ifndef HEADER
 #define HEADER
@@ -111,14 +111,14 @@
 // CHECK: call void @llvm.stackrestore(
 // CHECK: ret void
 
-// CHECK: define linkonce_odr{{.*}} void [[F_INT_LAMBDA]]([[CAP_TYPE2]]*
-// CHECK: [[THIS:%.+]] = load [[CAP_TYPE2]]*, [[CAP_TYPE2]]**
-// CHECK: [[SIZE_REF:%.+]] = getelementptr inbounds [[CAP_TYPE2]], [[CAP_TYPE2]]* [[THIS]], i{{.+}} 0, i{{.+}} 0
-// CHECK: [[SIZE:%.+]] = load [[INTPTR_T]], [[INTPTR_T]]* [[SIZE_REF]]
-// CHECK: call i{{.+}}* @llvm.stacksave()
-// CHECK: alloca [[INTPTR_T]], [[INTPTR_T]] [[SIZE]]
-// CHECK: call void @llvm.stackrestore(
-// CHECK: ret void
+// NOPCH: define linkonce_odr{{.*}} void [[F_INT_LAMBDA]]([[CAP_TYPE2]]*
+// NOPCH: [[THIS:%.+]] = load [[CAP_TYPE2]]*, [[CAP_TYPE2]]**
+// NOPCH: [[SIZE_REF:%.+]] = getelementptr inbounds [[CAP_TYPE2]], [[CAP_TYPE2]]* [[THIS]], i{{.+}} 0, i{{.+}} 0
+// NOPCH: [[SIZE:%.+]] = load [[INTPTR_T]], [[INTPTR_T]]* [[SIZE_REF]]
+// NOPCH: call i{{.+}}* @llvm.stacksave()
+// NOPCH: alloca [[INTPTR_T]], [[INTPTR_T]] [[SIZE]]
+// NOPCH: call void @llvm.stackrestore(
+// NOPCH: ret void
 
 // CHECK: define linkonce_odr{{.*}} void [[B_INT_LAMBDA]]([[CAP_TYPE3]]*
 // CHECK: [[SIZE2_REF:%.+]] = getelementptr inbounds [[CAP_TYPE3]], [[CAP_TYPE3]]* [[THIS:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
@@ -168,4 +168,14 @@
 // CHECK: [[MUL:%.+]] = mul {{.*}} i{{[0-9]+}} [[SIZE2]], [[SIZE1]]
 // CHECK: mul {{.*}} i{{[0-9]+}} {{[0-9]+}}, [[MUL]]
 // CHECK: ret void
+
+// PCH: define linkonce_odr{{.*}} void [[F_INT_LAMBDA]]([[CAP_TYPE2]]*
+// PCH: [[THIS:%.+]] = load [[CAP_TYPE2]]*, [[CAP_TYPE2]]**
+// PCH: [[SIZE_REF:%.+]] = getelementptr inbounds [[CAP_TYPE2]], [[CAP_TYPE2]]* [[THIS]], i{{.+}} 0, i{{.+}} 0
+// PCH: [[SIZE:%.+]] = load [[INTPTR_T]], [[INTPTR_T]]* [[SIZE_REF]]
+// PCH: call i{{.+}}* @llvm.stacksave()
+// PCH: alloca [[INTPTR_T]], [[INTPTR_T]] [[SIZE]]
+// PCH: call void @llvm.stackrestore(
+// PCH: ret void
+
 #endif
Index: clang/test/OpenMP/single_codegen.cpp
===================================================================
--- clang/test/OpenMP/single_codegen.cpp
+++ clang/test/OpenMP/single_codegen.cpp
@@ -1,6 +1,6 @@
-// RUN: %clang_cc1 -verify -fopenmp -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck --check-prefixes=CHECK,NOPCH %s
 // RUN: %clang_cc1 -fopenmp -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp -fnoopenmp-use-tls -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fnoopenmp-use-tls -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck --check-prefixes=CHECK,PCH %s
 // RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -std=c++11 -fopenmp -fnoopenmp-use-tls -fexceptions -fcxx-exceptions -debug-info-kind=line-tables-only -x c++ -emit-llvm %s -o - | FileCheck %s --check-prefix=TERM_DEBUG
 // RUN: %clang_cc1 -verify -fopenmp -fnoopenmp-use-tls -x c++ -std=c++11 -DARRAY -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=ARRAY %s
 
@@ -230,6 +230,58 @@
 // ARRAY: store %struct.St* %{{.+}}, %struct.St** %{{.+}},
 #endif
 
+// PCH-LABEL: @_ZN3SSTIdEC2Ev
+// PCH: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
+// PCH-NEXT: store double 0.000000e+00, double* %
+// PCH-NEXT: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
+// PCH-NEXT: store double* %{{.+}}, double** %
+// PCH-NEXT: load double*, double** %
+// PCH-NEXT: load double, double* %
+// PCH-NEXT: bitcast i64* %{{.+}} to double*
+// PCH-NEXT: store double %{{.+}}, double* %
+// PCH-NEXT: load i64, i64* %
+// PCH-NEXT: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* @{{.+}}, i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[SST_TY]]*, i64)* [[SST_MICROTASK:@.+]] to void
+// PCH-NEXT: ret void
+
+// PCH: define internal void [[SST_MICROTASK]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SST_TY]]* {{.+}}, i64 {{.+}})
+// PCH: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
+// PCH-NEXT: icmp ne i32 [[RES]], 0
+// PCH-NEXT: br i1
+
+// PCH: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
+// PCH-NEXT: load double*, double** %
+// PCH-NEXT: store double* %
+// PCH-LABEL: invoke void @_ZZN3SSTIdEC1EvENKUlvE_clEv(
+
+// PCH: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
+// PCH-NEXT: store i32 1, i32* [[DID_IT]],
+// PCH-NEXT: br label
+
+// PCH: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
+// PCH-NEXT: br label
+
+// PCH: getelementptr inbounds [1 x i8*], [1 x i8*]* [[LIST:%.+]], i64 0, i64 0
+// PCH: load double*, double** %
+// PCH-NEXT: bitcast double* %
+// PCH-NEXT: store i8* %
+// PCH-NEXT: bitcast [1 x i8*]* [[LIST]] to i8*
+// PCH-NEXT: load i32, i32* [[DID_IT]],
+// PCH-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 8, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
+// PCH-NEXT:  ret void
+
+// PCH-LABEL: @_ZZN3SSTIdEC1EvENKUlvE_clEv(
+// PCH: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
+// PCH-NEXT: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
+// PCH-NEXT: load double*, double** %
+// PCH-NEXT: store double* %
+// PCH-LABEL: call void @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
+// PCH-NEXT: ret void
+
+// PCH: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
+// PCH: ret void
+
+// PCH-LABEL: @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
+
 // CHECK-LABEL:@_ZN2SSC2ERi(
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* @{{.+}}, i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[SS_TY]]*, i64, i64, i64)* [[SS_MICROTASK:@.+]] to void
 // CHECK-NEXT: ret void
@@ -385,54 +437,54 @@
 // CHECK: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
 // CHECK: ret void
 
-// CHECK-LABEL: @_ZN3SSTIdEC2Ev
-// CHECK: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
-// CHECK-NEXT: store double 0.000000e+00, double* %
-// CHECK-NEXT: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
-// CHECK-NEXT: store double* %{{.+}}, double** %
-// CHECK-NEXT: load double*, double** %
-// CHECK-NEXT: load double, double* %
-// CHECK-NEXT: bitcast i64* %{{.+}} to double*
-// CHECK-NEXT: store double %{{.+}}, double* %
-// CHECK-NEXT: load i64, i64* %
-// CHECK-NEXT: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* @{{.+}}, i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[SST_TY]]*, i64)* [[SST_MICROTASK:@.+]] to void
-// CHECK-NEXT: ret void
-
-// CHECK: define internal void [[SST_MICROTASK]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SST_TY]]* {{.+}}, i64 {{.+}})
-// CHECK: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
-// CHECK-NEXT: icmp ne i32 [[RES]], 0
-// CHECK-NEXT: br i1
-
-// CHECK: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
-// CHECK-NEXT: load double*, double** %
-// CHECK-NEXT: store double* %
-// CHECK-LABEL: invoke void @_ZZN3SSTIdEC1EvENKUlvE_clEv(
-
-// CHECK: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
-// CHECK-NEXT: store i32 1, i32* [[DID_IT]],
-// CHECK-NEXT: br label
-
-// CHECK: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
-// CHECK-NEXT: br label
-
-// CHECK: getelementptr inbounds [1 x i8*], [1 x i8*]* [[LIST:%.+]], i64 0, i64 0
-// CHECK: load double*, double** %
-// CHECK-NEXT: bitcast double* %
-// CHECK-NEXT: store i8* %
-// CHECK-NEXT: bitcast [1 x i8*]* [[LIST]] to i8*
-// CHECK-NEXT: load i32, i32* [[DID_IT]],
-// CHECK-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 8, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
-// CHECK-NEXT:  ret void
-
-// CHECK-LABEL: @_ZZN3SSTIdEC1EvENKUlvE_clEv(
-// CHECK: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
-// CHECK-NEXT: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
-// CHECK-NEXT: load double*, double** %
-// CHECK-NEXT: store double* %
-// CHECK-LABEL: call void @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
-// CHECK-NEXT: ret void
-
-// CHECK: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
-// CHECK: ret void
-
-// CHECK-LABEL: @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
+// NOPCH-LABEL: @_ZN3SSTIdEC2Ev
+// NOPCH: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
+// NOPCH-NEXT: store double 0.000000e+00, double* %
+// NOPCH-NEXT: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
+// NOPCH-NEXT: store double* %{{.+}}, double** %
+// NOPCH-NEXT: load double*, double** %
+// NOPCH-NEXT: load double, double* %
+// NOPCH-NEXT: bitcast i64* %{{.+}} to double*
+// NOPCH-NEXT: store double %{{.+}}, double* %
+// NOPCH-NEXT: load i64, i64* %
+// NOPCH-NEXT: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* @{{.+}}, i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[SST_TY]]*, i64)* [[SST_MICROTASK:@.+]] to void
+// NOPCH-NEXT: ret void
+
+// NOPCH: define internal void [[SST_MICROTASK]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SST_TY]]* {{.+}}, i64 {{.+}})
+// NOPCH: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
+// NOPCH-NEXT: icmp ne i32 [[RES]], 0
+// NOPCH-NEXT: br i1
+
+// NOPCH: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
+// NOPCH-NEXT: load double*, double** %
+// NOPCH-NEXT: store double* %
+// NOPCH-LABEL: invoke void @_ZZN3SSTIdEC1EvENKUlvE_clEv(
+
+// NOPCH: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
+// NOPCH-NEXT: store i32 1, i32* [[DID_IT]],
+// NOPCH-NEXT: br label
+
+// NOPCH: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
+// NOPCH-NEXT: br label
+
+// NOPCH: getelementptr inbounds [1 x i8*], [1 x i8*]* [[LIST:%.+]], i64 0, i64 0
+// NOPCH: load double*, double** %
+// NOPCH-NEXT: bitcast double* %
+// NOPCH-NEXT: store i8* %
+// NOPCH-NEXT: bitcast [1 x i8*]* [[LIST]] to i8*
+// NOPCH-NEXT: load i32, i32* [[DID_IT]],
+// NOPCH-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 8, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
+// NOPCH-NEXT:  ret void
+
+// NOPCH-LABEL: @_ZZN3SSTIdEC1EvENKUlvE_clEv(
+// NOPCH: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
+// NOPCH-NEXT: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
+// NOPCH-NEXT: load double*, double** %
+// NOPCH-NEXT: store double* %
+// NOPCH-LABEL: call void @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
+// NOPCH-NEXT: ret void
+
+// NOPCH: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
+// NOPCH: ret void
+
+// NOPCH-LABEL: @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(