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/distribute_parallel_for_ast...

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

190 lines
8.1 KiB
C++
Raw Normal View History

[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
// RUN: %clang_cc1 -verify -std=c++11 -fopenmp -fopenmp-version=45 -ast-print %s -Wno-openmp-mapping | FileCheck %s --check-prefix=CHECK --check-prefix=OMP45
[OPENMP] Fix for PR31428: variable named like directive name modifier Directive name modifiers in 'if' clause are allowed only for OpenMP 4.5 and higher + in OpenMP 4.5 parsing procedure emits error message if ':' is not found after directive name modifier. llvm-svn: 290175
2016-12-20 20:10:05 +08:00
// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -emit-pch -o %t %s
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -std=c++11 -include-pch %t -fsyntax-only -verify %s -ast-print -Wno-openmp-mapping | FileCheck %s --check-prefix=CHECK --check-prefix=OMP45
// RUN: %clang_cc1 -verify -std=c++11 -fopenmp -fopenmp-version=50 -ast-print %s -Wno-openmp-mapping -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
// RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -x c++ -std=c++11 -emit-pch -o %t %s -DOMP5
// RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -std=c++11 -include-pch %t -fsyntax-only -verify %s -ast-print -Wno-openmp-mapping -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
[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
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
// RUN: %clang_cc1 -verify -std=c++11 -fopenmp-simd -fopenmp-version=45 -ast-print %s -Wno-openmp-mapping | FileCheck %s --check-prefix=CHECK --check-prefix=OMP45
[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 -fopenmp-version=45 -x c++ -std=c++11 -emit-pch -o %t %s
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
// RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=45 -std=c++11 -include-pch %t -fsyntax-only -verify %s -ast-print -Wno-openmp-mapping | FileCheck %s --check-prefix=CHECK --check-prefix=OMP45
// RUN: %clang_cc1 -verify -std=c++11 -fopenmp-simd -fopenmp-version=50 -ast-print %s -Wno-openmp-mapping -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
// RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=50 -x c++ -std=c++11 -emit-pch -o %t %s -DOMP5
// RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=50 -std=c++11 -include-pch %t -fsyntax-only -verify %s -ast-print -Wno-openmp-mapping -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
// expected-no-diagnostics
#ifndef HEADER
#define HEADER
struct S {
S(): a(0) {}
S(int v) : a(v) {}
int a;
typedef int type;
};
template <typename T>
class S7 : public T {
protected:
T a;
S7() : a(0) {}
public:
S7(typename T::type v) : a(v) {
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for private(a) private(this->a) private(T::a)
for (int k = 0; k < a.a; ++k)
++this->a.a;
}
S7 &operator=(S7 &s) {
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for private(a) private(this->a)
for (int k = 0; k < s.a.a; ++k)
++s.a.a;
return *this;
}
};
[OpenMP] Fix trailing space when printing pragmas, by Joel. E. Denny Summary: -ast-print prints omp pragmas with a trailing space. While this behavior is likely of little concern to most users, surely it's unintentional, and it's annoying for some source-level work I'm pursuing. This patch focuses on omp pragmas, but it also fixes init_seg and loop hint pragmas because they share implementation. The testing strategy here is to add usually just one '{{$}}' per relevant -ast-print test file. This seems to achieve good code coverage. However, this strategy is probably easy to forget as the tests evolve. That's probably fine as this fix is far from critical. The main goal of the testing is to aid the initial review. This patch also adds a fixme for "#pragma unroll", which prints as "#pragma unroll (enable)", which is invalid syntax. Reviewers: ABataev Reviewed By: ABataev Subscribers: guansong, cfe-commits Differential Revision: https://reviews.llvm.org/D43204 llvm-svn: 325145
2018-02-15 01:38:47 +08:00
// CHECK: #pragma omp distribute parallel for private(this->a) private(this->a) private(T::a){{$}}
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
// CHECK: #pragma omp distribute parallel for private(this->a) private(this->a)
Make output of -ast-print a valid C++ code. Output generated by option -ast-print looks like C/C++ code, and it really is for plain C. For C++ the produced output was not valid C++ code, but the differences were small. With this change the output is fixed and can be compiled. Tests are changed so that output produced by -ast-print is compiled again with the same flags and both outputs are compared. Option -ast-print is extensively used in clang tests but it itself was tested poorly, existing tests only checked that compiler did not crash. There are unit tests in file DeclPrinterTest.cpp, but they test only terse output mode. Differential Revision: https://reviews.llvm.org/D26452 llvm-svn: 286439
2016-11-10 16:49:37 +08:00
// CHECK: #pragma omp distribute parallel for private(this->a) private(this->a) private(this->S::a)
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
class S8 : public S7<S> {
S8() {}
public:
S8(int v) : S7<S>(v){
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for private(a) private(this->a) private(S7<S>::a)
for (int k = 0; k < a.a; ++k)
++this->a.a;
}
S8 &operator=(S8 &s) {
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for private(a) private(this->a)
for (int k = 0; k < s.a.a; ++k)
++s.a.a;
return *this;
}
};
// CHECK: #pragma omp distribute parallel for private(this->a) private(this->a) private(this->S7<S>::a)
// CHECK: #pragma omp distribute parallel for private(this->a) private(this->a)
template <class T, int N>
T tmain(T argc) {
T b = argc, c, d, e, f, h;
static T a;
// CHECK: static T a;
static T g;
#pragma omp threadprivate(g)
#pragma omp target
#pragma omp teams
[OPENMP]Initial support for 'allocate' clause. Added parsing/sema analysis of the allocate clause. llvm-svn: 357068
2019-03-27 22:14:31 +08:00
#pragma omp distribute parallel for dist_schedule(static, a) schedule(dynamic) default(none) copyin(g) firstprivate(a) allocate(a)
// CHECK: #pragma omp distribute parallel for dist_schedule(static, a) schedule(dynamic) default(none) copyin(g) firstprivate(a) allocate(a)
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
for (int i = 0; i < 2; ++i)
a = 2;
// CHECK-NEXT: for (int i = 0; i < 2; ++i)
// CHECK-NEXT: a = 2;
#pragma omp target
#pragma omp teams
[OPENMP]Initial support for 'allocate' clause. Added parsing/sema analysis of the allocate clause. llvm-svn: 357068
2019-03-27 22:14:31 +08:00
#pragma omp distribute parallel for allocate(argc) private(argc, b), firstprivate(c, d), lastprivate(f) collapse(N) schedule(static, N) if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h) dist_schedule(static,N)
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
for (int i = 0; i < 2; ++i)
for (int j = 0; j < 2; ++j)
for (int j = 0; j < 2; ++j)
for (int j = 0; j < 2; ++j)
for (int j = 0; j < 2; ++j)
for (int i = 0; i < 2; ++i)
for (int j = 0; j < 2; ++j)
for (int j = 0; j < 2; ++j)
for (int j = 0; j < 2; ++j)
for (int j = 0; j < 2; ++j)
[OPENMP] Do not allow variables to be first|last-privates in distribute directives. OpenMP standard does not allow to mark the variables as firstprivate and lastprivate at the same time in distribute-based directives. Patch fixes this problem. llvm-svn: 319560
2017-12-02 01:40:15 +08:00
a++;
[OPENMP]Initial support for 'allocate' clause. Added parsing/sema analysis of the allocate clause. llvm-svn: 357068
2019-03-27 22:14:31 +08:00
// CHECK: #pragma omp distribute parallel for allocate(argc) private(argc,b) firstprivate(c,d) lastprivate(f) collapse(N) schedule(static, N) if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h) dist_schedule(static, N)
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
// CHECK-NEXT: for (int i = 0; i < 2; ++i)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int i = 0; i < 2; ++i)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: for (int j = 0; j < 2; ++j)
// CHECK-NEXT: a++;
return T();
}
[OpenMP] Prepare sema to support combined constructs with omp distribute and omp for https://reviews.llvm.org/D32237 This patch prepares sema with additional fields to support all those composite and combined constructs of OpenMP that include pragma 'distribute' and 'for', such as 'distribute parallel for'. It also extends the regression tests for 'distribute parallel for' and adds a new one. llvm-svn: 300802
2017-04-20 08:39:39 +08:00
void foo(int argc, char **argv) {
int b, c, d, e, f, h;
static int a;
// CHECK: static int a;
static float g;
#pragma omp threadprivate(g)
[&]() {
#pragma omp target
#pragma omp teams
[OPENMP]Fix PR41767: diagnose DSA for variables in clauses with default(none). If the combined directive has default(none) clause and has clauses for inner directive that reference some variables, for which data-sharing attributes are not specified, the error messages should be emitted for such variables. llvm-svn: 360365
2019-05-10 02:44:53 +08:00
#pragma omp distribute parallel for schedule(guided, argc) default(none) copyin(g) dist_schedule(static, a) private(a) shared(argc)
// CHECK: #pragma omp distribute parallel for schedule(guided, argc) default(none) copyin(g) dist_schedule(static, a) private(a) shared(argc)
[OpenMP] Prepare sema to support combined constructs with omp distribute and omp for https://reviews.llvm.org/D32237 This patch prepares sema with additional fields to support all those composite and combined constructs of OpenMP that include pragma 'distribute' and 'for', such as 'distribute parallel for'. It also extends the regression tests for 'distribute parallel for' and adds a new one. llvm-svn: 300802
2017-04-20 08:39:39 +08:00
for (int i = 0; i < 2; ++i)
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
// CHECK: for (int i = 0; i < 2; ++i)
[OpenMP] Prepare sema to support combined constructs with omp distribute and omp for https://reviews.llvm.org/D32237 This patch prepares sema with additional fields to support all those composite and combined constructs of OpenMP that include pragma 'distribute' and 'for', such as 'distribute parallel for'. It also extends the regression tests for 'distribute parallel for' and adds a new one. llvm-svn: 300802
2017-04-20 08:39:39 +08:00
[&]() {
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
a = 2;
// CHECK: a = 2;
}();
[OpenMP] Prepare sema to support combined constructs with omp distribute and omp for https://reviews.llvm.org/D32237 This patch prepares sema with additional fields to support all those composite and combined constructs of OpenMP that include pragma 'distribute' and 'for', such as 'distribute parallel for'. It also extends the regression tests for 'distribute parallel for' and adds a new one. llvm-svn: 300802
2017-04-20 08:39:39 +08:00
}();
[&]() {
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for private(argc, b), firstprivate(argv, c), lastprivate(d, f) collapse(2) schedule(auto) if (argc) num_threads(a) default(shared) shared(e) reduction(+ : h) dist_schedule(static, b)
for (int i = 0; i < 10; ++i)
for (int j = 0; j < 10; ++j)
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
[&]() {
a++;
}();
[OpenMP] Prepare sema to support combined constructs with omp distribute and omp for https://reviews.llvm.org/D32237 This patch prepares sema with additional fields to support all those composite and combined constructs of OpenMP that include pragma 'distribute' and 'for', such as 'distribute parallel for'. It also extends the regression tests for 'distribute parallel for' and adds a new one. llvm-svn: 300802
2017-04-20 08:39:39 +08:00
// CHECK: #pragma omp distribute parallel for private(argc,b) firstprivate(argv,c) lastprivate(d,f) collapse(2) schedule(auto) if(argc) num_threads(a) default(shared) shared(e) reduction(+: h) dist_schedule(static, b)
// CHECK-NEXT: for (int i = 0; i < 10; ++i)
// CHECK-NEXT: for (int j = 0; j < 10; ++j)
// CHECK: a++;
}();
}
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
int main(int argc, char **argv) {
int b = argc, c, d, e, f, h;
static int a;
// CHECK: static int a;
static float g;
#pragma omp threadprivate(g)
#pragma omp target
#pragma omp teams
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
#ifdef OMP5
#pragma omp distribute parallel for schedule(guided, argc) default(none) copyin(g) dist_schedule(static, a) private(a) shared(argc) order(concurrent)
#else
[OPENMP]Fix PR41767: diagnose DSA for variables in clauses with default(none). If the combined directive has default(none) clause and has clauses for inner directive that reference some variables, for which data-sharing attributes are not specified, the error messages should be emitted for such variables. llvm-svn: 360365
2019-05-10 02:44:53 +08:00
#pragma omp distribute parallel for schedule(guided, argc) default(none) copyin(g) dist_schedule(static, a) private(a) shared(argc)
[OPENMP50]Basic parsing/sema analysis for order(concurrent) clause. Added parsing/sema/serialization support for order(concurrent) clause in loop|simd-based directives.
2020-02-01 05:09:26 +08:00
#endif // OMP5
// OMP45: #pragma omp distribute parallel for schedule(guided, argc) default(none) copyin(g) dist_schedule(static, a) private(a) shared(argc)
// OMP50: #pragma omp distribute parallel for schedule(guided, argc) default(none) copyin(g) dist_schedule(static, a) private(a) shared(argc) order(concurrent)
Resubmission of http://reviews.llvm.org/D21564 after fixes. [OpenMP] Initial implementation of parse and sema for composite pragma 'distribute parallel for' This patch is an initial implementation for #distribute parallel for. The main differences that affect other pragmas are: The implementation of 'distribute parallel for' requires blocking of the associated loop, where blocks are "distributed" to different teams and iterations within each block are scheduled to parallel threads within each team. To implement blocking, sema creates two additional worksharing directive fields that are used to pass the team assigned block lower and upper bounds through the outlined function resulting from 'parallel'. In this way, scheduling for 'for' to threads can use those bounds. As a consequence of blocking, the stride of 'distribute' is not 1 but it is equal to the blocking size. This is returned by the runtime and sema prepares a DistIncrExpr variable to hold that value. As a consequence of blocking, the global upper bound (EnsureUpperBound) expression of the 'for' is not the original loop upper bound (e.g. in for(i = 0 ; i < N; i++) this is 'N') but it is the team-assigned block upper bound. Sema creates a new expression holding the calculation of the actual upper bound for 'for' as UB = min(UB, PrevUB), where UB is the loop upper bound, and PrevUB is the team-assigned block upper bound. llvm-svn: 273884
2016-06-27 22:55:37 +08:00
for (int i = 0; i < 2; ++i)
a = 2;
// CHECK-NEXT: for (int i = 0; i < 2; ++i)
// CHECK-NEXT: a = 2;
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for private(argc, b), firstprivate(argv, c), lastprivate(d, f) collapse(2) schedule(auto) if (argc) num_threads(a) default(shared) shared(e) reduction(+ : h) dist_schedule(static, b)
for (int i = 0; i < 10; ++i)
for (int j = 0; j < 10; ++j)
a++;
// CHECK: #pragma omp distribute parallel for private(argc,b) firstprivate(argv,c) lastprivate(d,f) collapse(2) schedule(auto) if(argc) num_threads(a) default(shared) shared(e) reduction(+: h) dist_schedule(static, b)
// CHECK-NEXT: for (int i = 0; i < 10; ++i)
// CHECK-NEXT: for (int j = 0; j < 10; ++j)
// CHECK-NEXT: a++;
return (tmain<int, 5>(argc) + tmain<char, 1>(argv[0][0]));
}
#endif