maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/test/OpenMP/single_codegen.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

497 lines
24 KiB
C++
Raw Normal View History

[OpenMP] Upgrade default version of OpenMP to 5.0 Summary: When -fopenmp option is specified then version 5.0 will be set as default. Reviewers: gregrodgers, jdoerfert, ABataev Reviewed By: ABataev Subscribers: pdhaliwal, yaxunl, guansong, sstefan1, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D81098
2020-06-25 15:01:15 +08:00
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=50 -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s --check-prefixes=OMP50,CHECK
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=50 -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -fopenmp-version=45 -o - | FileCheck %s --check-prefixes=OMP45,CHECK
// RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -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 -fopenmp-version=50 -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 --check-prefixes=OMP50,CHECK
// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -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 -fopenmp-version=45 -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 --check-prefixes=OMP45,CHECK
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// 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
[OpenMP] Add TLS-based implementation for threadprivate directive. llvm-svn: 242080
2015-07-14 06:54:53 +08:00
// 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
[OPENMP] Support for -fopenmp-simd option with compilation of simd loops only. Added support for -fopenmp-simd option that allows compilation of simd-based constructs without emission of OpenMP runtime calls. llvm-svn: 321560
2017-12-30 02:07:07 +08:00
Revert "[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition" This reverts commit 789a46f2d742e11edaade28cb59a0f4d2a1d770e. Accidentally committed.
2020-02-04 02:09:39 +08:00
// RUN: %clang_cc1 -verify -fopenmp-simd -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck --check-prefix SIMD-ONLY0 %s
[OPENMP] Support for -fopenmp-simd option with compilation of simd loops only. Added support for -fopenmp-simd option that allows compilation of simd-based constructs without emission of OpenMP runtime calls. llvm-svn: 321560
2017-12-30 02:07:07 +08:00
// RUN: %clang_cc1 -fopenmp-simd -fnoopenmp-use-tls -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s
Revert "[CodeGenModule] Assume dso_local for -fpic -fno-semantic-interposition" This reverts commit 789a46f2d742e11edaade28cb59a0f4d2a1d770e. Accidentally committed.
2020-02-04 02:09:39 +08:00
// RUN: %clang_cc1 -fopenmp-simd -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-prefix SIMD-ONLY0 %s
[OPENMP] Support for -fopenmp-simd option with compilation of simd loops only. Added support for -fopenmp-simd option that allows compilation of simd-based constructs without emission of OpenMP runtime calls. llvm-svn: 321560
2017-12-30 02:07:07 +08:00
// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -std=c++11 -fopenmp-simd -fnoopenmp-use-tls -fexceptions -fcxx-exceptions -debug-info-kind=line-tables-only -x c++ -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
// RUN: %clang_cc1 -verify -fopenmp-simd -fnoopenmp-use-tls -x c++ -std=c++11 -DARRAY -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
// SIMD-ONLY0-NOT: {{__kmpc|__tgt}}
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
// expected-no-diagnostics
[OPENMP] Fixed codegen for copying/initialization of array variables/parameters. This modification generates proper copyin/initialization sequences for array variables/parameters. Before they were considered as pointers, not arrays. llvm-svn: 237691
2015-05-19 20:31:28 +08:00
#ifndef ARRAY
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
#ifndef HEADER
#define HEADER
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
class TestClass {
public:
int a;
TestClass() : a(0) {}
TestClass(const TestClass &C) : a(C.a) {}
TestClass &operator=(const TestClass &) { return *this;}
~TestClass(){};
};
// CHECK-DAG: [[TEST_CLASS_TY:%.+]] = type { i{{[0-9]+}} }
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-DAG: [[SST_TY:%.+]] = type { double }
// CHECK-DAG: [[SS_TY:%.+]] = type { i32, i8, i32* }
// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
[OpenMP][NFC] Reuse OMPIRBuilder `struct ident_t` handling in Clang Replace the `ident_t` handling in Clang with the methods offered by the OMPIRBuilder. This cuts down on the clang code as well as the differences between the two, making further transitions easier. Tests have changed but there should not be a real functional change. The most interesting difference is probably that we stop generating local ident_t allocations for now and just use globals. Given that this happens only with debug info, the location part of the `ident_t` is probably bigger than the test anyway. As the location part is already a global, we can avoid the allocation, memcpy, and store in favor of a constant global that is slightly bigger. This can be revisited if there are complications. Reviewed By: ABataev Differential Revision: https://reviews.llvm.org/D80735
2020-05-28 22:41:01 +08:00
// CHECK: [[IMPLICIT_BARRIER_SINGLE_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 322, i32 0, i32 0, i8*
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
// CHECK: define void [[FOO:@.+]]()
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
TestClass tc;
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
TestClass tc2[2];
#pragma omp threadprivate(tc, tc2)
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
[CodeGen] Emit a call instruction instead of an invoke if the called llvm function is marked nounwind This fixes cases where an invoke is emitted, despite the called llvm function being marked nounwind, because ConstructAttributeList failed to add the attribute to the attribute list. llvm optimization passes turn invokes into calls and optimize away the exception handling code, but it's better to avoid emitting the code in the front-end if the called function is known not to raise an exception. Differential Revision: https://reviews.llvm.org/D83906
2020-07-16 05:47:45 +08:00
void foo() { extern void mayThrow(); mayThrow(); }
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
struct SS {
int a;
int b : 4;
int &c;
SS(int &d) : a(0), b(0), c(d) {
#pragma omp parallel firstprivate(a, b, c)
#pragma omp single copyprivate(a, this->b, (this)->c)
[&]() {
++this->a, --b, (this)->c /= 1;
#pragma omp parallel firstprivate(a, b, c)
#pragma omp single copyprivate(a, this->b, (this)->c)
++(this)->a, --b, this->c /= 1;
}();
}
};
template<typename T>
struct SST {
T a;
SST() : a(T()) {
#pragma omp parallel firstprivate(a)
#pragma omp single copyprivate(this->a)
[&]() {
[&]() {
++this->a;
#pragma omp parallel firstprivate(a)
#pragma omp single copyprivate((this)->a)
++(this)->a;
}();
}();
}
};
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
// CHECK-LABEL: @main
[OPENMP] Improved code for generating debug info + generation of all OpenMP regions in termination scope Patch adds proper generation of debug info for all OpenMP regions. Also, all OpenMP regions are generated in a termination scope, because standard does not allow to throw exceptions out of structured blocks, associated with the OpenMP regions Differential Revision: http://reviews.llvm.org/D7935 llvm-svn: 231757
2015-03-10 13:15:26 +08:00
// TERM_DEBUG-LABEL: @main
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
int main() {
[OPENMP] Fix emission of the __kmpc_global_thread_num. Fixed emission of the __kmpc_global_thread_num() so that it is not messed up with alloca instructions anymore. Plus, fixes emission of the __kmpc_global_thread_num() functions in the target outlined regions so that they are not called before runtime is initialized. llvm-svn: 343856
2018-10-05 23:08:53 +08:00
// CHECK: alloca i32
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK-DAG: [[A_ADDR:%.+]] = alloca i8
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK-DAG: [[A2_ADDR:%.+]] = alloca [2 x i8]
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK-DAG: [[C_ADDR:%.+]] = alloca [[TEST_CLASS_TY]]
[OPENMP] Fix emission of the __kmpc_global_thread_num. Fixed emission of the __kmpc_global_thread_num() so that it is not messed up with alloca instructions anymore. Plus, fixes emission of the __kmpc_global_thread_num() functions in the target outlined regions so that they are not called before runtime is initialized. llvm-svn: 343856
2018-10-05 23:08:53 +08:00
// CHECK-DAG: [[DID_IT:%.+]] = alloca i32,
// CHECK-DAG: [[COPY_LIST:%.+]] = alloca [5 x i8*],
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
char a;
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
char a2[2];
[OPENMP] Fix for http://llvm.org/PR24674: assertion failed and and abort trap Fix processing of shared variables with reference types in OpenMP constructs. Previously, if the variable was not marked in one of the private clauses, the reference to this variable was emitted incorrectly and caused an assertion later. llvm-svn: 246846
2015-09-04 19:26:21 +08:00
TestClass &c = tc;
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
SST<double> sst;
SS ss(c.a);
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
// CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T_TY]]* [[DEFAULT_LOC:@.+]])
// CHECK: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
// CHECK-NEXT: [[IS_SINGLE:%.+]] = icmp ne i32 [[RES]], 0
// CHECK-NEXT: br i1 [[IS_SINGLE]], label {{%?}}[[THEN:.+]], label {{%?}}[[EXIT:.+]]
// CHECK: [[THEN]]
// CHECK-NEXT: store i8 2, i8* [[A_ADDR]]
// CHECK-NEXT: call void @__kmpc_end_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
// CHECK-NEXT: br label {{%?}}[[EXIT]]
// CHECK: [[EXIT]]
[OPENMP] Do not emit implicit barrier for single directive with 'copyprivate' clause(s). Runtime function for 'copyprivate' directive generates implicit barriers, so no need to emit it. Differential Revision: http://reviews.llvm.org/D9215 llvm-svn: 235692
2015-04-24 12:00:39 +08:00
// CHECK-NOT: call {{.+}} @__kmpc_cancel_barrier
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
#pragma omp single nowait
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
a = 2;
// CHECK: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
// CHECK-NEXT: [[IS_SINGLE:%.+]] = icmp ne i32 [[RES]], 0
// CHECK-NEXT: br i1 [[IS_SINGLE]], label {{%?}}[[THEN:.+]], label {{%?}}[[EXIT:.+]]
// CHECK: [[THEN]]
[OPENMP] Do not emit implicit barrier for single directive with 'copyprivate' clause(s). Runtime function for 'copyprivate' directive generates implicit barriers, so no need to emit it. Differential Revision: http://reviews.llvm.org/D9215 llvm-svn: 235692
2015-04-24 12:00:39 +08:00
// CHECK-NEXT: store i8 2, i8* [[A_ADDR]]
// CHECK-NEXT: call void @__kmpc_end_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
// CHECK-NEXT: br label {{%?}}[[EXIT]]
// CHECK: [[EXIT]]
[OPENMP 4.0] Fixed codegen for 'cancellation point' construct. Generate the next code for 'cancellation point': if (__kmpc_cancellationpoint()) { __kmpc_cancel_barrier(); <exit construct>; } llvm-svn: 241336
2015-07-03 17:56:58 +08:00
// CHECK: call{{.*}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_SINGLE_LOC]], i32 [[GTID]])
[OPENMP] Do not emit implicit barrier for single directive with 'copyprivate' clause(s). Runtime function for 'copyprivate' directive generates implicit barriers, so no need to emit it. Differential Revision: http://reviews.llvm.org/D9215 llvm-svn: 235692
2015-04-24 12:00:39 +08:00
#pragma omp single
a = 2;
[OPENMP] Fixed codegen for 'copyprivate' clause. Fixed initialization of 'single' region completion + changed type of the third argument of __kmpc_copyprivate() runtime function to size_t. llvm-svn: 236198
2015-04-30 11:47:32 +08:00
// CHECK: store i32 0, i32* [[DID_IT]]
[OPENMP] Do not emit implicit barrier for single directive with 'copyprivate' clause(s). Runtime function for 'copyprivate' directive generates implicit barriers, so no need to emit it. Differential Revision: http://reviews.llvm.org/D9215 llvm-svn: 235692
2015-04-24 12:00:39 +08:00
// CHECK: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
// CHECK-NEXT: [[IS_SINGLE:%.+]] = icmp ne i32 [[RES]], 0
// CHECK-NEXT: br i1 [[IS_SINGLE]], label {{%?}}[[THEN:.+]], label {{%?}}[[EXIT:.+]]
// CHECK: [[THEN]]
[OPENMP] Improved code for generating debug info + generation of all OpenMP regions in termination scope Patch adds proper generation of debug info for all OpenMP regions. Also, all OpenMP regions are generated in a termination scope, because standard does not allow to throw exceptions out of structured blocks, associated with the OpenMP regions Differential Revision: http://reviews.llvm.org/D7935 llvm-svn: 231757
2015-03-10 13:15:26 +08:00
// CHECK-NEXT: invoke void [[FOO]]()
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: to label {{%?}}[[CONT:.+]] unwind
// CHECK: [[CONT]]
Revert "[OPENMP] Allow runtime insert its own code inside OpenMP regions." Reverting because of failed tests. llvm-svn: 264577
2016-03-28 20:58:34 +08:00
// CHECK: call void @__kmpc_end_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
[OPENMP] Allow runtime insert its own code inside OpenMP regions. Solution unifies interface of RegionCodeGenTy type to allow insert runtime-specific code before/after main codegen action defined in CGStmtOpenMP.cpp file. Runtime should not define its own RegionCodeGenTy for general OpenMP directives, but must be allowed to insert its own (required) code to support target specific codegen. llvm-svn: 264700
2016-03-29 13:34:15 +08:00
// CHECK: store i32 1, i32* [[DID_IT]]
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
// CHECK-NEXT: br label {{%?}}[[EXIT]]
// CHECK: [[EXIT]]
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[A_PTR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[COPY_LIST]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: store i8* [[A_ADDR]], i8** [[A_PTR_REF]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[C_PTR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[COPY_LIST]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
[OPENMP] Fix for http://llvm.org/PR24674: assertion failed and and abort trap Fix processing of shared variables with reference types in OpenMP constructs. Previously, if the variable was not marked in one of the private clauses, the reference to this variable was emitted incorrectly and caused an assertion later. llvm-svn: 246846
2015-09-04 19:26:21 +08:00
// CHECK: store i8* {{.+}}, i8** [[C_PTR_REF]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[TC_PTR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[COPY_LIST]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: [[TC_THREADPRIVATE_ADDR_VOID_PTR:%.+]] = call{{.*}} i8* @__kmpc_threadprivate_cached
// CHECK: [[TC_THREADPRIVATE_ADDR:%.+]] = bitcast i8* [[TC_THREADPRIVATE_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
// CHECK: [[TC_PTR_REF_VOID_PTR:%.+]] = bitcast [[TEST_CLASS_TY]]* [[TC_THREADPRIVATE_ADDR]] to i8*
// CHECK: store i8* [[TC_PTR_REF_VOID_PTR]], i8** [[TC_PTR_REF]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[A2_PTR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[COPY_LIST]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
// CHECK: [[BITCAST:%.+]] = bitcast [2 x i8]* [[A2_ADDR]] to i8*
// CHECK: store i8* [[BITCAST]], i8** [[A2_PTR_REF]],
// CHECK: [[TC2_PTR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[COPY_LIST]], i{{[0-9]+}} 0, i{{[0-9]+}} 4
// CHECK: [[TC2_THREADPRIVATE_ADDR_VOID_PTR:%.+]] = call{{.*}} i8* @__kmpc_threadprivate_cached
// CHECK: [[TC2_THREADPRIVATE_ADDR:%.+]] = bitcast i8* [[TC2_THREADPRIVATE_ADDR_VOID_PTR]] to [2 x [[TEST_CLASS_TY]]]*
// CHECK: [[TC2_PTR_REF_VOID_PTR:%.+]] = bitcast [2 x [[TEST_CLASS_TY]]]* [[TC2_THREADPRIVATE_ADDR]] to i8*
// CHECK: store i8* [[TC2_PTR_REF_VOID_PTR]], i8** [[TC2_PTR_REF]],
// CHECK: [[COPY_LIST_VOID_PTR:%.+]] = bitcast [5 x i8*]* [[COPY_LIST]] to i8*
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: [[DID_IT_VAL:%.+]] = load i32, i32* [[DID_IT]],
[OPENMP] Fixed codegen for 'copyprivate' clause. Fixed initialization of 'single' region completion + changed type of the third argument of __kmpc_copyprivate() runtime function to size_t. llvm-svn: 236198
2015-04-30 11:47:32 +08:00
// CHECK: call void @__kmpc_copyprivate([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i64 40, i8* [[COPY_LIST_VOID_PTR]], void (i8*, i8*)* [[COPY_FUNC:@.+]], i32 [[DID_IT_VAL]])
[OPENMP] Do not emit implicit barrier for single directive with 'copyprivate' clause(s). Runtime function for 'copyprivate' directive generates implicit barriers, so no need to emit it. Differential Revision: http://reviews.llvm.org/D9215 llvm-svn: 235692
2015-04-24 12:00:39 +08:00
// CHECK-NOT: call {{.+}} @__kmpc_cancel_barrier
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
#pragma omp single copyprivate(a, c, tc, a2, tc2)
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
foo();
// CHECK-NOT: call i32 @__kmpc_single
// CHECK-NOT: call void @__kmpc_end_single
return a;
}
[OpenMP] Upgrade default version of OpenMP to 5.0 Summary: When -fopenmp option is specified then version 5.0 will be set as default. Reviewers: gregrodgers, jdoerfert, ABataev Reviewed By: ABataev Subscribers: pdhaliwal, yaxunl, guansong, sstefan1, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D81098
2020-06-25 15:01:15 +08:00
// OMP50-LABEL: declare i8* @__kmpc_threadprivate_cached(
IR: print value numbers for unnamed function arguments For consistency with normal instructions and clarity when reading IR, it's best to print the %0, %1, ... names of function arguments in definitions. Also modifies the parser to accept IR in that form for obvious reasons. llvm-svn: 367755
2019-08-03 22:28:34 +08:00
// CHECK: void [[COPY_FUNC]](i8* %0, i8* %1)
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: store i8* %0, i8** [[DST_ADDR_REF:%.+]],
// CHECK: store i8* %1, i8** [[SRC_ADDR_REF:%.+]],
// CHECK: [[DST_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[DST_ADDR_REF]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[DST_ADDR:%.+]] = bitcast i8* [[DST_ADDR_VOID_PTR]] to [5 x i8*]*
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: [[SRC_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[SRC_ADDR_REF]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[SRC_ADDR:%.+]] = bitcast i8* [[SRC_ADDR_VOID_PTR]] to [5 x i8*]*
// CHECK: [[DST_A_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[DST_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: [[DST_A_ADDR:%.+]] = load i8*, i8** [[DST_A_ADDR_REF]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[SRC_A_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[SRC_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
// CHECK: [[SRC_A_ADDR:%.+]] = load i8*, i8** [[SRC_A_ADDR_REF]],
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: [[SRC_A_VAL:%.+]] = load i8, i8* [[SRC_A_ADDR]],
// CHECK: store i8 [[SRC_A_VAL]], i8* [[DST_A_ADDR]],
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[DST_C_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[DST_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
// CHECK: [[DST_C_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[DST_C_ADDR_REF]],
// CHECK: [[DST_C_ADDR:%.+]] = bitcast i8* [[DST_C_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
// CHECK: [[SRC_C_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[SRC_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
// CHECK: [[SRC_C_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[SRC_C_ADDR_REF]],
// CHECK: [[SRC_C_ADDR:%.+]] = bitcast i8* [[SRC_C_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: call{{.*}} [[TEST_CLASS_TY_ASSIGN:@.+]]([[TEST_CLASS_TY]]* [[DST_C_ADDR]], [[TEST_CLASS_TY]]* {{.*}}[[SRC_C_ADDR]])
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[DST_TC_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[DST_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
// CHECK: [[DST_TC_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[DST_TC_ADDR_REF]],
// CHECK: [[DST_TC_ADDR:%.+]] = bitcast i8* [[DST_TC_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
// CHECK: [[SRC_TC_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[SRC_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
// CHECK: [[SRC_TC_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[SRC_TC_ADDR_REF]],
// CHECK: [[SRC_TC_ADDR:%.+]] = bitcast i8* [[SRC_TC_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
// CHECK: call{{.*}} [[TEST_CLASS_TY_ASSIGN]]([[TEST_CLASS_TY]]* [[DST_TC_ADDR]], [[TEST_CLASS_TY]]* {{.*}}[[SRC_TC_ADDR]])
// CHECK: [[DST_A2_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[DST_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
// CHECK: [[DST_A2_ADDR:%.+]] = load i8*, i8** [[DST_A2_ADDR_REF]],
// CHECK: [[SRC_A2_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[SRC_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
// CHECK: [[SRC_A2_ADDR:%.+]] = load i8*, i8** [[SRC_A2_ADDR_REF]],
Change memcpy/memove/memset to have dest and source alignment attributes (Step 1). Summary: Upstream LLVM is changing the the prototypes of the @llvm.memcpy/memmove/memset intrinsics. This change updates the Clang tests for this change. The @llvm.memcpy/memmove/memset intrinsics currently have an explicit argument which is required to be a constant integer. It represents the alignment of the dest (and source), and so must be the minimum of the actual alignment of the two. This change removes the alignment argument in favour of placing the alignment attribute on the source and destination pointers of the memory intrinsic call. For example, code which used to read: call void @llvm.memcpy.p0i8.p0i8.i32(i8* %dest, i8* %src, i32 100, i32 4, i1 false) will now read call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %dest, i8* align 4 %src, i32 100, i1 false) At this time the source and destination alignments must be the same (Step 1). Step 2 of the change, to be landed shortly, will relax that contraint and allow the source and destination to have different alignments. llvm-svn: 322964
2018-01-20 01:12:54 +08:00
// CHECK: call void @llvm.memcpy.{{.+}}(i8* align 1 [[DST_A2_ADDR]], i8* align 1 [[SRC_A2_ADDR]], i{{[0-9]+}} 2, i1 false)
[OPENMP] Fixed codegen for arrays in 'copyprivate' clause. Fixed a bug with codegen of variables with array types specified in 'copyprivate' clause of 'single' directive. Differential Revision: http://reviews.llvm.org/D8914 llvm-svn: 234856
2015-04-14 13:11:24 +08:00
// CHECK: [[DST_TC2_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[DST_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 4
// CHECK: [[DST_TC2_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[DST_TC2_ADDR_REF]],
// CHECK: [[DST_TC2_ADDR:%.+]] = bitcast i8* [[DST_TC2_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
// CHECK: [[SRC_TC2_ADDR_REF:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[SRC_ADDR]], i{{[0-9]+}} 0, i{{[0-9]+}} 4
// CHECK: [[SRC_TC2_ADDR_VOID_PTR:%.+]] = load i8*, i8** [[SRC_TC2_ADDR_REF]],
// CHECK: [[SRC_TC2_ADDR:%.+]] = bitcast i8* [[SRC_TC2_ADDR_VOID_PTR]] to [[TEST_CLASS_TY]]*
// CHECK: br i1
// CHECK: call{{.*}} [[TEST_CLASS_TY_ASSIGN]]([[TEST_CLASS_TY]]* %{{.+}}, [[TEST_CLASS_TY]]* {{.*}})
// CHECK: br i1
[OPENMP] Codegen for 'copyprivate' clause ('single' directive). If there is at least one 'copyprivate' clause is associated with the single directive, the following code is generated: ``` i32 did_it = 0; \\ for 'copyprivate' clause if(__kmpc_single(ident_t *, gtid)) { SingleOpGen(); __kmpc_end_single(ident_t *, gtid); did_it = 1; \\ for 'copyprivate' clause } <copyprivate_list>[0] = &var0; ... <copyprivate_list>[n] = &varn; call __kmpc_copyprivate(ident_t *, gtid, <copyprivate_list_size>, <copyprivate_list>, <copy_func>, did_it); ... void<copy_func>(void *LHSArg, void *RHSArg) { Dst = (void * [n])(LHSArg); Src = (void * [n])(RHSArg); Dst[0] = Src[0]; ... Dst[n] = Src[n]; } ``` All list items from all 'copyprivate' clauses are gathered into single <copyprivate list> (<copyprivate_list_size> is a size in bytes of this list) and <copy_func> is used to propagate values of private or threadprivate variables from the 'single' region to other implicit threads from outer 'parallel' region. Differential Revision: http://reviews.llvm.org/D8410 llvm-svn: 232932
2015-03-23 14:18:07 +08:00
// CHECK: ret void
[OpenMP] Upgrade default version of OpenMP to 5.0 Summary: When -fopenmp option is specified then version 5.0 will be set as default. Reviewers: gregrodgers, jdoerfert, ABataev Reviewed By: ABataev Subscribers: pdhaliwal, yaxunl, guansong, sstefan1, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D81098
2020-06-25 15:01:15 +08:00
// OMP50-LABEL: void @_ZN3SSTIdEC2Ev(
// OMP50: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
// OMP50-NEXT: store double 0.000000e+00, double* %
// OMP50-NEXT: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
// OMP50-NEXT: store double* %{{.+}}, double** %
// OMP50-NEXT: load double*, double** %
// OMP50-NEXT: load double, double* %
// OMP50-NEXT: bitcast i64* %{{.+}} to double*
// OMP50-NEXT: store double %{{.+}}, double* %
// OMP50-NEXT: load i64, i64* %
// OMP50-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
// OMP50-NEXT: ret void
// OMP50: define internal void [[SST_MICROTASK]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SST_TY]]* {{.+}}, i64 {{.+}})
// OMP50: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// OMP50-NEXT: icmp ne i32 [[RES]], 0
// OMP50-NEXT: br i1
// OMP50: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
// OMP50-NEXT: load double*, double** %
// OMP50-NEXT: store double* %
// OMP50-LABEL: invoke void @_ZZN3SSTIdEC1EvENKUlvE_clEv(
// OMP50: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// OMP50-NEXT: store i32 1, i32* [[DID_IT]],
// OMP50-NEXT: br label
// OMP50: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// OMP50-NEXT: br label
// OMP50: getelementptr inbounds [1 x i8*], [1 x i8*]* [[LIST:%.+]], i64 0, i64 0
// OMP50: load double*, double** %
// OMP50-NEXT: bitcast double* %
// OMP50-NEXT: store i8* %
// OMP50-NEXT: bitcast [1 x i8*]* [[LIST]] to i8*
// OMP50-NEXT: load i32, i32* [[DID_IT]],
// OMP50-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 8, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
// OMP50-NEXT: ret void
// OMP50-LABEL: @_ZZN3SSTIdEC1EvENKUlvE_clEv(
// OMP50: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
// OMP50-NEXT: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
// OMP50-NEXT: load double*, double** %
// OMP50-NEXT: store double* %
// OMP50-LABEL: call void @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
// OMP50-NEXT: ret void
// OMP50: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
// OMP50: ret void
// OMP45-LABEL: parallel_single
[OPENMP] Improved code for generating debug info + generation of all OpenMP regions in termination scope Patch adds proper generation of debug info for all OpenMP regions. Also, all OpenMP regions are generated in a termination scope, because standard does not allow to throw exceptions out of structured blocks, associated with the OpenMP regions Differential Revision: http://reviews.llvm.org/D7935 llvm-svn: 231757
2015-03-10 13:15:26 +08:00
// TERM_DEBUG-LABEL: parallel_single
void parallel_single() {
[OPENMP] Fixed codegen for directives without function outlining. Fixed crash on codegen for directives like 'omp for', 'omp single' etc. inside of the 'omp parallel', 'omp task' etc. regions. llvm-svn: 230621
2015-02-26 18:27:34 +08:00
#pragma omp parallel
#pragma omp single
[OPENMP] Fix PR35156: Get correct thread id with windows exceptions. If the thread id is requested in windows mode within funclets, we may generate incorrect function call that could lead to broken codegen. llvm-svn: 317208
2017-11-02 22:25:34 +08:00
// TERM_DEBUG-NOT: __kmpc_global_thread_num
[OPENMP] Improved code for generating debug info + generation of all OpenMP regions in termination scope Patch adds proper generation of debug info for all OpenMP regions. Also, all OpenMP regions are generated in a termination scope, because standard does not allow to throw exceptions out of structured blocks, associated with the OpenMP regions Differential Revision: http://reviews.llvm.org/D7935 llvm-svn: 231757
2015-03-10 13:15:26 +08:00
// TERM_DEBUG: call i32 @__kmpc_single({{.+}}), !dbg [[DBG_LOC_START:![0-9]+]]
// TERM_DEBUG: invoke void {{.*}}foo{{.*}}()
// TERM_DEBUG: unwind label %[[TERM_LPAD:.+]],
// TERM_DEBUG-NOT: __kmpc_global_thread_num
// TERM_DEBUG: call void @__kmpc_end_single({{.+}}), !dbg [[DBG_LOC_END:![0-9]+]]
[OPENMP] Fixed tests for non-debug builds. NFC. llvm-svn: 231758
2015-03-10 13:28:46 +08:00
// TERM_DEBUG: [[TERM_LPAD]]
[OPENMP] Improved code for generating debug info + generation of all OpenMP regions in termination scope Patch adds proper generation of debug info for all OpenMP regions. Also, all OpenMP regions are generated in a termination scope, because standard does not allow to throw exceptions out of structured blocks, associated with the OpenMP regions Differential Revision: http://reviews.llvm.org/D7935 llvm-svn: 231757
2015-03-10 13:15:26 +08:00
// TERM_DEBUG: call void @__clang_call_terminate
// TERM_DEBUG: unreachable
foo();
[OPENMP] Fixed codegen for directives without function outlining. Fixed crash on codegen for directives like 'omp for', 'omp single' etc. inside of the 'omp parallel', 'omp task' etc. regions. llvm-svn: 230621
2015-02-26 18:27:34 +08:00
}
DebugInfo: Metadata constructs now start with DI* LLVM r236120 renamed debug info IR constructs to use a `DI` prefix, now that the `DIDescriptor` hierarchy has been gone for about a week. This commit was generated using the rename-md-di-nodes.sh upgrade script attached to PR23080, followed by running clang-format-diff.py on the `lib/` portion of the patch. llvm-svn: 236121
2015-04-30 00:40:08 +08:00
// TERM_DEBUG-DAG: [[DBG_LOC_START]] = !DILocation(line: [[@LINE-12]],
// TERM_DEBUG-DAG: [[DBG_LOC_END]] = !DILocation(line: [[@LINE-3]],
[OPENMP] Fixed codegen for copying/initialization of array variables/parameters. This modification generates proper copyin/initialization sequences for array variables/parameters. Before they were considered as pointers, not arrays. llvm-svn: 237691
2015-05-19 20:31:28 +08:00
#endif
#else
// ARRAY-LABEL: array_func
struct St {
int a, b;
St() : a(0), b(0) {}
St &operator=(const St &) { return *this; };
~St() {}
};
[OPENMP] Fixed codegen for directives without function outlining. Fixed crash on codegen for directives like 'omp for', 'omp single' etc. inside of the 'omp parallel', 'omp task' etc. regions. llvm-svn: 230621
2015-02-26 18:27:34 +08:00
[OPENMP] Fixed codegen for parameters privatization. For parameters we shall take a derived type of parameters, not the original one. llvm-svn: 237882
2015-05-21 17:47:46 +08:00
void array_func(int n, int a[n], St s[2]) {
[OPENMP] General code improvements. llvm-svn: 330140
2018-04-17 01:59:34 +08:00
// ARRAY: call void @__kmpc_copyprivate(%struct.ident_t* @{{.+}}, i32 %{{.+}}, i64 16, i8* %{{.+}}, void (i8*, i8*)* [[CPY:@.+]], i32 %{{.+}})
[OPENMP] Fixed codegen for copying/initialization of array variables/parameters. This modification generates proper copyin/initialization sequences for array variables/parameters. Before they were considered as pointers, not arrays. llvm-svn: 237691
2015-05-19 20:31:28 +08:00
#pragma omp single copyprivate(a, s)
;
}
// ARRAY: define internal void [[CPY]]
[OPENMP] Fixed codegen for parameters privatization. For parameters we shall take a derived type of parameters, not the original one. llvm-svn: 237882
2015-05-21 17:47:46 +08:00
// ARRAY: store i32* %{{.+}}, i32** %{{.+}},
// ARRAY: store %struct.St* %{{.+}}, %struct.St** %{{.+}},
[OPENMP] Initial codegen for 'single' directive. This patch emits the following code for the single directive: #pragma omp single <body> <----> if(__kmpc_single(...)) { <body> __kmpc_end_single(...); } Differential Revision: http://reviews.llvm.org/D7045 llvm-svn: 228275
2015-02-05 14:35:41 +08:00
#endif
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-LABEL:@_ZN2SSC2ERi(
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// 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
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-NEXT: ret void
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: define internal void [[SS_MICROTASK]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SS_TY]]* {{.+}}, i64 {{.+}}, i64 {{.+}}, i64 {{.+}})
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// Private a
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: alloca i64,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// Private b
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: alloca i64,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// Private c
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: alloca i64,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: alloca i32*,
[OPENMP] Pass scalar firstprivate vars by value. For better performance and to unify code with offloading part we pass scalar firstprivate values by value, instead of by reference. It will remove some extra copying operations. llvm-svn: 269751
2016-05-17 16:55:33 +08:00
// CHECK: alloca i32*,
// CHECK: alloca i32*,
// CHECK: alloca i32*,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: [[DID_IT:%.+]] = alloca i32,
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: bitcast i64* %{{.+}} to i32*
// CHECK: bitcast i64* %{{.+}} to i32*
// CHECK: bitcast i64* %{{.+}} to i32*
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: store i32 0, i32* [[DID_IT]],
// CHECK: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// CHECK-NEXT: icmp ne i32 [[RES]], 0
// CHECK-NEXT: br i1
// CHECK: getelementptr inbounds [[CAP_TY:%.+]], [[CAP_TY]]* [[CAP:%.+]], i32 0, i32 0
// CHECK: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 1
// CHECK-NEXT: load i32*, i32** %
// CHECK-NEXT: store i32* %
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 2
// CHECK-NEXT: store i32* %
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 3
// CHECK-NEXT: load i32*, i32** %
// CHECK-NEXT: store i32* %
// CHECK-LABEL: invoke void @_ZZN2SSC1ERiENKUlvE_clEv(
// CHECK-SAME: [[CAP_TY]]* [[CAP]])
Revert "[OPENMP] Allow runtime insert its own code inside OpenMP regions." Reverting because of failed tests. llvm-svn: 264577
2016-03-28 20:58:34 +08:00
// CHECK: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
[OPENMP] Allow runtime insert its own code inside OpenMP regions. Solution unifies interface of RegionCodeGenTy type to allow insert runtime-specific code before/after main codegen action defined in CGStmtOpenMP.cpp file. Runtime should not define its own RegionCodeGenTy for general OpenMP directives, but must be allowed to insert its own (required) code to support target specific codegen. llvm-svn: 264700
2016-03-29 13:34:15 +08:00
// CHECK: store i32 1, i32* [[DID_IT]],
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: br label
// CHECK: call void @__kmpc_end_single(%{{.+}}* @{{.+}}, i32 %{{.+}})
// CHECK: br label
// CHECK: getelementptr inbounds [3 x i8*], [3 x i8*]* [[LIST:%.+]], i64 0, i64 0
// CHECK: load i32*, i32** %
// CHECK-NEXT: bitcast i32* %
// CHECK-NEXT: store i8* %
// CHECK: getelementptr inbounds [3 x i8*], [3 x i8*]* [[LIST]], i64 0, i64 1
// CHECK-NEXT: bitcast i32* %
// CHECK-NEXT: store i8* %
// CHECK: getelementptr inbounds [3 x i8*], [3 x i8*]* [[LIST]], i64 0, i64 2
// CHECK: load i32*, i32** %
// CHECK-NEXT: bitcast i32* %
// CHECK-NEXT: store i8* %
// CHECK-NEXT: bitcast [3 x i8*]* [[LIST]] to i8*
// CHECK-NEXT: load i32, i32* [[DID_IT]],
// CHECK-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 24, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
// CHECK-NEXT: ret void
// CHECK-LABEL: @_ZZN2SSC1ERiENKUlvE_clEv(
// CHECK: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP:%.+]], i32 0, i32 1
// CHECK-NEXT: load i32*, i32** %
// CHECK-NEXT: load i32, i32* %
// CHECK-NEXT: add nsw i32 %{{.+}}, 1
// CHECK-NEXT: store i32 %
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 2
// CHECK-NEXT: load i32*, i32** %
// CHECK-NEXT: load i32, i32* %
// CHECK-NEXT: add nsw i32 %{{.+}}, -1
// CHECK-NEXT: store i32 %
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 3
// CHECK-NEXT: load i32*, i32** %
// CHECK-NEXT: load i32, i32* %
// CHECK-NEXT: sdiv i32 %{{.+}}, 1
// CHECK-NEXT: store i32 %
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 1
// CHECK-NEXT: load i32*, i32** %
[OPENMP] Pass scalar firstprivate vars by value. For better performance and to unify code with offloading part we pass scalar firstprivate values by value, instead of by reference. It will remove some extra copying operations. llvm-svn: 269751
2016-05-17 16:55:33 +08:00
// CHECK-NEXT: load i32, i32* %
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK-NEXT: bitcast i64* %
// CHECK-NEXT: store i32 %{{.+}}, i32* %
// CHECK-NEXT: load i64, i64* %
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 2
// CHECK-NEXT: load i32*, i32** %
[OPENMP] Pass scalar firstprivate vars by value. For better performance and to unify code with offloading part we pass scalar firstprivate values by value, instead of by reference. It will remove some extra copying operations. llvm-svn: 269751
2016-05-17 16:55:33 +08:00
// CHECK-NEXT: load i32, i32* %
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK-NEXT: bitcast i64* %
// CHECK-NEXT: store i32 %{{.+}}, i32* %
// CHECK-NEXT: load i64, i64* %
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-NEXT: getelementptr inbounds [[CAP_TY]], [[CAP_TY]]* [[CAP]], i32 0, i32 3
// CHECK-NEXT: load i32*, i32** %
[OPENMP] Pass scalar firstprivate vars by value. For better performance and to unify code with offloading part we pass scalar firstprivate values by value, instead of by reference. It will remove some extra copying operations. llvm-svn: 269751
2016-05-17 16:55:33 +08:00
// CHECK-NEXT: load i32, i32* %
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK-NEXT: bitcast i64* %
// CHECK-NEXT: store i32 %{{.+}}, i32* %
// CHECK-NEXT: load i64, i64* %
// CHECK-NEXT: 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_MICROTASK1:@.+]] to void
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-NEXT: ret void
IR: print value numbers for unnamed function arguments For consistency with normal instructions and clarity when reading IR, it's best to print the %0, %1, ... names of function arguments in definitions. Also modifies the parser to accept IR in that form for obvious reasons. llvm-svn: 367755
2019-08-03 22:28:34 +08:00
// CHECK: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: ret void
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: define internal void [[SS_MICROTASK1]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SS_TY]]* {{.+}}, i64 {{.+}}, i64 {{.+}}, i64 {{.+}})
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// Private a
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: alloca i64,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// Private b
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: alloca i64,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// Private c
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: alloca i64,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: alloca i32*,
[OPENMP] Pass scalar firstprivate vars by value. For better performance and to unify code with offloading part we pass scalar firstprivate values by value, instead of by reference. It will remove some extra copying operations. llvm-svn: 269751
2016-05-17 16:55:33 +08:00
// CHECK: alloca i32*,
// CHECK: alloca i32*,
// CHECK: alloca i32*,
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: [[DID_IT:%.+]] = alloca i32,
Re-apply r272900 - [OpenMP] Cast captures by copy when passed to fork call so that they are compatible to what the runtime library expects. An issue in one of the regression tests was fixed for 32-bit hosts. llvm-svn: 272931
2016-06-17 02:39:34 +08:00
// CHECK: bitcast i64* %{{.+}} to i32*
// CHECK: bitcast i64* %{{.+}} to i32*
// CHECK: bitcast i64* %{{.+}} to i32*
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// CHECK-NEXT: icmp ne i32 [[RES]], 0
// CHECK-NEXT: br i1
// CHECK-NOT: getelementptr inbounds
// CHECK: load i32*, i32** %
// CHECK-NEXT: load i32, i32* %
// CHECK-NEXT: add nsw i32 %{{.+}}, 1
// CHECK-NEXT: store i32 %
// CHECK-NOT: getelementptr inbounds
// CHECK: load i32, i32* %
// CHECK-NEXT: add nsw i32 %{{.+}}, -1
// CHECK-NEXT: store i32 %
// CHECK-NOT: getelementptr inbounds
// CHECK: load i32*, i32** %
// CHECK-NEXT: load i32, i32* %
// CHECK-NEXT: sdiv i32 %{{.+}}, 1
// CHECK-NEXT: store i32 %
Revert "[OPENMP] Allow runtime insert its own code inside OpenMP regions." Reverting because of failed tests. llvm-svn: 264577
2016-03-28 20:58:34 +08:00
// CHECK-NEXT: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
[OPENMP] Allow runtime insert its own code inside OpenMP regions. Solution unifies interface of RegionCodeGenTy type to allow insert runtime-specific code before/after main codegen action defined in CGStmtOpenMP.cpp file. Runtime should not define its own RegionCodeGenTy for general OpenMP directives, but must be allowed to insert its own (required) code to support target specific codegen. llvm-svn: 264700
2016-03-29 13:34:15 +08:00
// CHECK-NEXT: store i32 1, i32* [[DID_IT]],
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK-NEXT: br label
// CHECK: getelementptr inbounds [3 x i8*], [3 x i8*]* [[LIST:%.+]], i64 0, i64 0
// CHECK: load i32*, i32** %
// CHECK-NEXT: bitcast i32* %
// CHECK-NEXT: store i8* %
// CHECK: getelementptr inbounds [3 x i8*], [3 x i8*]* [[LIST]], i64 0, i64 1
// CHECK-NEXT: bitcast i32* %
// CHECK-NEXT: store i8* %
// CHECK: getelementptr inbounds [3 x i8*], [3 x i8*]* [[LIST]], i64 0, i64 2
// CHECK: load i32*, i32** %
// CHECK-NEXT: bitcast i32* %
// CHECK-NEXT: store i8* %
// CHECK-NEXT: bitcast [3 x i8*]* [[LIST]] to i8*
// CHECK-NEXT: load i32, i32* [[DID_IT]],
// CHECK-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 24, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
// CHECK-NEXT: ret void
IR: print value numbers for unnamed function arguments For consistency with normal instructions and clarity when reading IR, it's best to print the %0, %1, ... names of function arguments in definitions. Also modifies the parser to accept IR in that form for obvious reasons. llvm-svn: 367755
2019-08-03 22:28:34 +08:00
// CHECK: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
[OPENMP 4.5] Allow to use private data members in 'copyprivate' clause. OpenMP 4.5 allows privatization of non-static data members in non-static member functions. This patch adds support of private data members in 'copyprivate' clauses. llvm-svn: 263706
2016-03-17 18:50:17 +08:00
// CHECK: ret void
[OpenMP] Upgrade default version of OpenMP to 5.0 Summary: When -fopenmp option is specified then version 5.0 will be set as default. Reviewers: gregrodgers, jdoerfert, ABataev Reviewed By: ABataev Subscribers: pdhaliwal, yaxunl, guansong, sstefan1, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D81098
2020-06-25 15:01:15 +08:00
// OMP45-LABEL: void @_ZN3SSTIdEC2Ev(
// OMP45: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
// OMP45-NEXT: store double 0.000000e+00, double* %
// OMP45-NEXT: getelementptr inbounds [[SST_TY]], [[SST_TY]]* %{{.+}}, i32 0, i32 0
// OMP45-NEXT: store double* %{{.+}}, double** %
// OMP45-NEXT: load double*, double** %
// OMP45-NEXT: load double, double* %
// OMP45-NEXT: bitcast i64* %{{.+}} to double*
// OMP45-NEXT: store double %{{.+}}, double* %
// OMP45-NEXT: load i64, i64* %
// OMP45-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
// OMP45-NEXT: ret void
// OMP45: define internal void [[SST_MICROTASK]](i32* {{[^,]+}}, i32* {{[^,]+}}, [[SST_TY]]* {{.+}}, i64 {{.+}})
// OMP45: [[RES:%.+]] = call i32 @__kmpc_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// OMP45-NEXT: icmp ne i32 [[RES]], 0
// OMP45-NEXT: br i1
// OMP45: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
// OMP45-NEXT: load double*, double** %
// OMP45-NEXT: store double* %
// OMP45-LABEL: invoke void @_ZZN3SSTIdEC1EvENKUlvE_clEv(
// OMP45: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// OMP45-NEXT: store i32 1, i32* [[DID_IT]],
// OMP45-NEXT: br label
// OMP45: call void @__kmpc_end_single([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}})
// OMP45-NEXT: br label
// OMP45: getelementptr inbounds [1 x i8*], [1 x i8*]* [[LIST:%.+]], i64 0, i64 0
// OMP45: load double*, double** %
// OMP45-NEXT: bitcast double* %
// OMP45-NEXT: store i8* %
// OMP45-NEXT: bitcast [1 x i8*]* [[LIST]] to i8*
// OMP45-NEXT: load i32, i32* [[DID_IT]],
// OMP45-NEXT: call void @__kmpc_copyprivate([[IDENT_T_TY]]* @{{.+}}, i32 %{{.+}}, i64 8, i8* %{{.+}}, void (i8*, i8*)* [[COPY_FUNC:@[^,]+]], i32 %{{.+}})
// OMP45-NEXT: ret void
// OMP45-LABEL: @_ZZN3SSTIdEC1EvENKUlvE_clEv(
// OMP45: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
// OMP45-NEXT: getelementptr inbounds %{{.+}}, %{{.+}}* %{{.+}}, i32 0, i32 1
// OMP45-NEXT: load double*, double** %
// OMP45-NEXT: store double* %
// OMP45-LABEL: call void @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(
// OMP45-NEXT: ret void
// OMP45: define internal void [[COPY_FUNC]](i8* %0, i8* %1)
// OMP45: ret void
// OMP45-LABEL: @_ZZZN3SSTIdEC1EvENKUlvE_clEvENKUlvE_clEv(