diff options
| author | Alexey Bataev <a.bataev@hotmail.com> | 2015-05-07 04:25:17 +0000 |
|---|---|---|
| committer | Alexey Bataev <a.bataev@hotmail.com> | 2015-05-07 04:25:17 +0000 |
| commit | 53223c986cfbc16179dc82353812038ce3362900 (patch) | |
| tree | 231373c939510933031b4aa8f165d32d317d50a9 /clang/test/OpenMP/for_codegen.cpp | |
| parent | 4fce73aabf7f17ce9a43ba55d7158bc3da129487 (diff) | |
| download | bcm5719-llvm-53223c986cfbc16179dc82353812038ce3362900.tar.gz bcm5719-llvm-53223c986cfbc16179dc82353812038ce3362900.zip | |
[OPENMP] Generate !llvm.mem.loop_parallel_access metadata for loops with dynamic/guided scheduling.
Inner bodies of OpenMP worksharing loop-based constructs with dynamic or guided scheduling are allowed to be marked with !llvm.mem.parallel_loop_access metadata for better optimization. Worksharing constructs with static scheduling cannot be marked this way (according to OpenMP standard "A data dependence between the same logical iterations in two such loops is guaranteed").
Constructs with auto and runtime scheduling are also not marked because automatically chosen scheduling may be static also.
Differential Revision: http://reviews.llvm.org/D9518
llvm-svn: 236693
Diffstat (limited to 'clang/test/OpenMP/for_codegen.cpp')
| -rw-r--r-- | clang/test/OpenMP/for_codegen.cpp | 9 |
1 files changed, 7 insertions, 2 deletions
diff --git a/clang/test/OpenMP/for_codegen.cpp b/clang/test/OpenMP/for_codegen.cpp index 31161e26e1e..6939108933c 100644 --- a/clang/test/OpenMP/for_codegen.cpp +++ b/clang/test/OpenMP/for_codegen.cpp @@ -41,6 +41,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) { // ... loop body ... // End of body: store into a[i]: // CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK-NOT: !llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i32, i32* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add nsw i32 [[IV1_2]], 1 @@ -81,6 +82,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) { // ... loop body ... // End of body: store into a[i]: // CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK-NOT: !llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i32, i32* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add nsw i32 [[IV1_2]], 1 @@ -129,6 +131,7 @@ void static_chunked(float *a, float *b, float *c, float *d) { // ... loop body ... // End of body: store into a[i]: // CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK-NOT: !llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i32, i32* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add i32 [[IV1_2]], 1 @@ -180,7 +183,7 @@ void dynamic1(float *a, float *b, float *c, float *d) { // CHECK-NEXT: store i64 [[CALC_I_2]], i64* [[LC_I:.+]] // ... loop body ... // End of body: store into a[i]: -// CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK: store float [[RESULT:%.+]], float* {{%.+}}!llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i64, i64* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add i64 [[IV1_2]], 1 @@ -221,7 +224,7 @@ void guided7(float *a, float *b, float *c, float *d) { // CHECK-NEXT: store i64 [[CALC_I_2]], i64* [[LC_I:.+]] // ... loop body ... // End of body: store into a[i]: -// CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK: store float [[RESULT:%.+]], float* {{%.+}}!llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i64, i64* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add i64 [[IV1_2]], 1 @@ -266,6 +269,7 @@ void test_auto(float *a, float *b, float *c, float *d) { // ... loop body ... // End of body: store into a[i]: // CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK-NOT: !llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i64, i64* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add nsw i64 [[IV1_2]], 1 @@ -307,6 +311,7 @@ void runtime(float *a, float *b, float *c, float *d) { // ... loop body ... // End of body: store into a[i]: // CHECK: store float [[RESULT:%.+]], float* {{%.+}} +// CHECK-NOT: !llvm.mem.parallel_loop_access a[i] = b[i] * c[i] * d[i]; // CHECK: [[IV1_2:%.+]] = load i32, i32* [[OMP_IV]]{{.*}} // CHECK-NEXT: [[ADD1_2:%.+]] = add nsw i32 [[IV1_2]], 1 |
