From: Gheorghe-Teodor Bercea <gheorghe-teod.bercea@ibm.com>
Date: Thu, 29 Nov 2018 20:53:49 +0000 (+0000)
Subject: [OpenMP] Add a new version of the SPMD deinit kernel function
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=e1fea8ec605dd8710ffd680c51610f211da4b50c;p=clang

[OpenMP] Add a new version of the SPMD deinit kernel function

Summary: This patch adds a new runtime for the SPMD deinit kernel function which replaces the previous function. The new function takes as argument the flag which signals whether the runtime is required or not. This enables the compiler to optimize out the part of the deinit function which are not needed.

Reviewers: ABataev, caomhin

Reviewed By: ABataev

Subscribers: jholewinski, guansong, cfe-commits

Differential Revision: https://reviews.llvm.org/D54970

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@347915 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp b/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
index def3ba7c41..d4b599eeff 100644
--- a/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
+++ b/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
@@ -33,8 +33,8 @@ enum OpenMPRTLFunctionNVPTX {
   /// Call to void __kmpc_spmd_kernel_init(kmp_int32 thread_limit,
   /// int16_t RequiresOMPRuntime, int16_t RequiresDataSharing);
   OMPRTL_NVPTX__kmpc_spmd_kernel_init,
-  /// Call to void __kmpc_spmd_kernel_deinit();
-  OMPRTL_NVPTX__kmpc_spmd_kernel_deinit,
+  /// Call to void __kmpc_spmd_kernel_deinit_v2(int16_t RequiresOMPRuntime);
+  OMPRTL_NVPTX__kmpc_spmd_kernel_deinit_v2,
   /// Call to void __kmpc_kernel_prepare_parallel(void
   /// *outlined_function, int16_t
   /// IsOMPRuntimeInitialized);
@@ -1413,8 +1413,11 @@ void CGOpenMPRuntimeNVPTX::emitSPMDEntryFooter(CodeGenFunction &CGF,
 
   CGF.EmitBlock(OMPDeInitBB);
   // DeInitialize the OMP state in the runtime; called by all active threads.
+  llvm::Value *Args[] = {/*RequiresOMPRuntime=*/
+                         CGF.Builder.getInt16(RequiresFullRuntime ? 1 : 0)};
   CGF.EmitRuntimeCall(
-      createNVPTXRuntimeFunction(OMPRTL_NVPTX__kmpc_spmd_kernel_deinit), None);
+      createNVPTXRuntimeFunction(
+          OMPRTL_NVPTX__kmpc_spmd_kernel_deinit_v2), Args);
   CGF.EmitBranch(EST.ExitBB);
 
   CGF.EmitBlock(EST.ExitBB);
@@ -1597,11 +1600,12 @@ CGOpenMPRuntimeNVPTX::createNVPTXRuntimeFunction(unsigned Function) {
     RTLFn = CGM.CreateRuntimeFunction(FnTy, "__kmpc_spmd_kernel_init");
     break;
   }
-  case OMPRTL_NVPTX__kmpc_spmd_kernel_deinit: {
-    // Build void __kmpc_spmd_kernel_deinit();
+  case OMPRTL_NVPTX__kmpc_spmd_kernel_deinit_v2: {
+    // Build void __kmpc_spmd_kernel_deinit_v2(int16_t RequiresOMPRuntime);
+    llvm::Type *TypeParams[] = {CGM.Int16Ty};
     auto *FnTy =
-        llvm::FunctionType::get(CGM.VoidTy, llvm::None, /*isVarArg*/ false);
-    RTLFn = CGM.CreateRuntimeFunction(FnTy, "__kmpc_spmd_kernel_deinit");
+        llvm::FunctionType::get(CGM.VoidTy, TypeParams, /*isVarArg*/ false);
+    RTLFn = CGM.CreateRuntimeFunction(FnTy, "__kmpc_spmd_kernel_deinit_v2");
     break;
   }
   case OMPRTL_NVPTX__kmpc_kernel_prepare_parallel: {
diff --git a/test/OpenMP/nvptx_target_parallel_codegen.cpp b/test/OpenMP/nvptx_target_parallel_codegen.cpp
index bbde7bcff9..7964d768d7 100644
--- a/test/OpenMP/nvptx_target_parallel_codegen.cpp
+++ b/test/OpenMP/nvptx_target_parallel_codegen.cpp
@@ -68,7 +68,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[DONE:.+]]
   //
   // CHECK: [[DONE]]
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   // CHECK: br label {{%?}}[[EXIT:.+]]
   //
   // CHECK: [[EXIT]]
@@ -111,7 +111,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[DONE:.+]]
   //
   // CHECK: [[DONE]]
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   // CHECK: br label {{%?}}[[EXIT:.+]]
   //
   // CHECK: [[EXIT]]
diff --git a/test/OpenMP/nvptx_target_parallel_proc_bind_codegen.cpp b/test/OpenMP/nvptx_target_parallel_proc_bind_codegen.cpp
index 905487fed7..4b32f9d87b 100644
--- a/test/OpenMP/nvptx_target_parallel_proc_bind_codegen.cpp
+++ b/test/OpenMP/nvptx_target_parallel_proc_bind_codegen.cpp
@@ -57,7 +57,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[DONE:.+]]
   //
   // CHECK: [[DONE]]
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   // CHECK: br label {{%?}}[[EXIT:.+]]
   //
   // CHECK: [[EXIT]]
@@ -79,7 +79,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[DONE:.+]]
   //
   // CHECK: [[DONE]]
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   // CHECK: br label {{%?}}[[EXIT:.+]]
   //
   // CHECK: [[EXIT]]
@@ -100,7 +100,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[DONE:.+]]
   //
   // CHECK: [[DONE]]
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   // CHECK: br label {{%?}}[[EXIT:.+]]
   //
   // CHECK: [[EXIT]]
diff --git a/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp b/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
index 1687c8ea76..965ff9e4fb 100644
--- a/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
+++ b/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
@@ -60,7 +60,7 @@ int bar(int n){
   //
   // CHECK: [[EXECUTE]]
   // CHECK: {{call|invoke}} void [[PFN:@.+]](i32*
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   //
   //
   // define internal void [[PFN]](
@@ -254,7 +254,7 @@ int bar(int n){
   //
   // CHECK: [[EXECUTE]]
   // CHECK: {{call|invoke}} void [[PFN1:@.+]](i32*
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   //
   //
   // define internal void [[PFN1]](
@@ -527,7 +527,7 @@ int bar(int n){
   //
   // CHECK: [[EXECUTE]]
   // CHECK: {{call|invoke}} void [[PFN2:@.+]](i32*
-  // CHECK: call void @__kmpc_spmd_kernel_deinit()
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
   //
   //
   // define internal void [[PFN2]](
diff --git a/test/OpenMP/nvptx_target_teams_codegen.cpp b/test/OpenMP/nvptx_target_teams_codegen.cpp
index c62d254524..33c8b065d3 100644
--- a/test/OpenMP/nvptx_target_teams_codegen.cpp
+++ b/test/OpenMP/nvptx_target_teams_codegen.cpp
@@ -234,7 +234,7 @@ int bar(int n){
 // CHECK: call void [[L0:@.+]](i32* %{{.+}}, i32* %{{.+}}, i16* %{{.*}})
 // CHECK-NOT: call void @__kmpc_end_serialized_parallel(
 // CHECK-NOT: call void @__kmpc_data_sharing_pop_stack(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
 // CHECK: ret
 
 // CHECK: define internal void [[L0]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i16* dereferenceable
diff --git a/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp b/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp
index 79315717ae..0c06a60e44 100644
--- a/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp
+++ b/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp
@@ -84,7 +84,7 @@ int bar(int n){
 // CHECK: {{call|invoke}} void [[OUTL1:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
 // CHECK: call void @__kmpc_restore_team_static_memory(i16 1)
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL1]](
@@ -98,7 +98,7 @@ int bar(int n){
 // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
 // CHECK: {{call|invoke}} void [[OUTL2:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL2]](
@@ -112,7 +112,7 @@ int bar(int n){
 // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
 // CHECK: {{call|invoke}} void [[OUTL3:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL3]](
@@ -204,7 +204,7 @@ int bar(int n){
 
 // CHECK: [[DIST_INNER_LOOP_END]]:
 // CHECK: call void @__kmpc_for_static_fini(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL4]](
diff --git a/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp b/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp
index 9464475325..f7ce262961 100644
--- a/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp
+++ b/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp
@@ -32,7 +32,7 @@ int main(int argc, char **argv) {
 
 // CHECK: call void @__kmpc_for_static_fini(%struct.ident_t* @
 
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 
 // CHECK: define internal void [[PARALLEL]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i{{64|32}} %{{.+}}, i{{64|32}} %{{.+}}, i{{64|32}} [[ARGC:%.+]], i32* dereferenceable{{.*}})
 // CHECK-NOT: call i8* @__kmpc_data_sharing_push_stack(
diff --git a/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp b/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp
index 2e9ceb129b..5a828af687 100644
--- a/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp
+++ b/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp
@@ -79,7 +79,7 @@ int bar(int n){
 // CHECK: {{call|invoke}} void [[OUTL1:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
 // CHECK: call void @__kmpc_restore_team_static_memory(i16 1)
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL1]](
@@ -93,7 +93,7 @@ int bar(int n){
 // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
 // CHECK: {{call|invoke}} void [[OUTL2:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL2]](
@@ -107,7 +107,7 @@ int bar(int n){
 // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
 // CHECK: {{call|invoke}} void [[OUTL3:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL3]](
@@ -123,7 +123,7 @@ int bar(int n){
 // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91, {{.+}}, {{.+}}, {{.+}}* [[COMB_UB]],
 // CHECK: {{call|invoke}} void [[OUTL4:@.+]](
 // CHECK: call void @__kmpc_for_static_fini(
-// CHECK: call void @__kmpc_spmd_kernel_deinit()
+// CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 0)
 // CHECK: ret void
 
 // CHECK: define internal void [[OUTL4]](
diff --git a/test/OpenMP/nvptx_teams_reduction_codegen.cpp b/test/OpenMP/nvptx_teams_reduction_codegen.cpp
index d0a31b0238..97a128866a 100644
--- a/test/OpenMP/nvptx_teams_reduction_codegen.cpp
+++ b/test/OpenMP/nvptx_teams_reduction_codegen.cpp
@@ -98,6 +98,7 @@ int bar(int n){
   // CHECK: define {{.*}}void [[T2:@__omp_offloading_.+template.+l43]](
   //
   // CHECK: {{call|invoke}} void [[T2]]_worker()
+
   //
   // CHECK: call void @__kmpc_kernel_init(
   //
@@ -176,7 +177,7 @@ int bar(int n){
   //
   // CHECK: [[EXIT]]
   // call void @__kmpc_restore_team_static_memory(i16 1)
-  // CHECK: call void @__kmpc_spmd_kernel_deinit(
+  // CHECK: call void @__kmpc_spmd_kernel_deinit_v2(i16 1)
 
   // CHECK: define internal void [[OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i32* dereferenceable{{.+}}, i16* dereferenceable{{.+}})
   //
@@ -286,7 +287,6 @@ int bar(int n){
   // CHECK: [[MAXV:%.+]] = phi i16 [ [[MAX1]], %[[DO_MAX]] ], [ [[MAX2]], %[[MAX_ELSE]] ]
   // CHECK: store i16 [[MAXV]], i16* [[VAR2_LHS]],
   // CHECK: ret void
-
   //
   // Shuffle and reduce function
   // CHECK: define internal void [[PAR_SHUFFLE_REDUCE_FN]](i8*, i16 {{.*}}, i16 {{.*}}, i16 {{.*}})