mirror of
https://github.com/llvm/llvm-project.git
synced 2025-04-25 13:36:08 +00:00

Given "loop" construct, clang will try to treat it as "for", "distribute" or "simd", depending on either the implied binding, or the bind clause if present. This patch moves the code that performs this construct remapping from sema to codegen. For a "loop" construct without a bind clause, this patch will create an implicit bind clause based on implied binding to simplify further analysis. During codegen the function `EmitOMPGenericLoopDirective` (i.e. "loop") will invoke the "emit" functions for "for", "distribute" or "simd", depending on the bind clause. --------- Co-authored-by: Alexey Bataev <a.bataev@gmx.com>
149 lines
3.3 KiB
C++
149 lines
3.3 KiB
C++
// RUN: %clang_cc1 -fopenmp -emit-pch -o %t.a %s
|
|
// RUN: %clang_cc1 -fopenmp -include-pch %t.a %s -ast-print -o - | FileCheck %s
|
|
|
|
// CHECK: #pragma clang loop vectorize_width(4)
|
|
// CHECK: #pragma clang loop interleave_count(8)
|
|
// CHECK: #pragma clang loop unroll_count(16){{$}}
|
|
// CHECK: #pragma clang loop vectorize(enable)
|
|
// CHECK: #pragma clang loop interleave(disable)
|
|
// CHECK: #pragma clang loop unroll(disable)
|
|
// CHECK: #pragma clang loop distribute(enable)
|
|
// CHECK: #pragma clang loop vectorize(disable)
|
|
// CHECK: #pragma clang loop interleave(enable)
|
|
// CHECK: #pragma clang loop unroll(full)
|
|
// CHECK: #pragma clang loop distribute(disable)
|
|
// FIXME: "#pragma unroll (enable)" is invalid and is not the input source.
|
|
// CHECK: #pragma unroll (enable){{$}}
|
|
// CHECK: #pragma unroll (32){{$}}
|
|
// CHECK: #pragma nounroll{{$}}
|
|
// CHECK: #pragma clang loop vectorize_width(V)
|
|
// CHECK: #pragma clang loop interleave_count(I)
|
|
// CHECK: #pragma omp loop bind(thread)
|
|
// CHECK: #pragma omp loop bind(parallel)
|
|
// CHECK: #pragma omp loop bind(teams)
|
|
|
|
#ifndef HEADER
|
|
#define HEADER
|
|
|
|
class pragma_test {
|
|
public:
|
|
inline void run1(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma clang loop vectorize_width(4)
|
|
#pragma clang loop interleave_count(8)
|
|
#pragma clang loop unroll_count(16)
|
|
while (i < Length) {
|
|
List[i] = i;
|
|
i++;
|
|
}
|
|
}
|
|
|
|
inline void run2(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma clang loop vectorize(enable)
|
|
#pragma clang loop interleave(disable)
|
|
#pragma clang loop unroll(disable)
|
|
#pragma clang loop distribute(enable)
|
|
while (i - 1 < Length) {
|
|
List[i] = i;
|
|
i++;
|
|
}
|
|
}
|
|
|
|
inline void run3(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma clang loop vectorize(disable)
|
|
#pragma clang loop interleave(enable)
|
|
#pragma clang loop unroll(full)
|
|
#pragma clang loop distribute(disable)
|
|
while (i - 3 < Length) {
|
|
List[i] = i;
|
|
i++;
|
|
}
|
|
}
|
|
|
|
inline void run4(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma unroll
|
|
while (i - 3 < Length) {
|
|
List[i] = i;
|
|
i++;
|
|
}
|
|
}
|
|
|
|
inline void run5(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma unroll 32
|
|
while (i - 3 < Length) {
|
|
List[i] = i;
|
|
i++;
|
|
}
|
|
}
|
|
|
|
inline void run6(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma nounroll
|
|
while (i - 3 < Length) {
|
|
List[i] = i;
|
|
i++;
|
|
}
|
|
}
|
|
|
|
template <int V, int I>
|
|
inline void run7(int *List, int Length) {
|
|
#pragma clang loop vectorize_width(V)
|
|
#pragma clang loop interleave_count(I)
|
|
for (int i = 0; i < Length; i++) {
|
|
List[i] = i;
|
|
}
|
|
}
|
|
|
|
inline void run8(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma omp loop bind(thread)
|
|
for (int i = 0; i < Length; i++) {
|
|
List[i] = i;
|
|
}
|
|
}
|
|
|
|
inline void run9(int *List, int Length) {
|
|
int i = 0;
|
|
#pragma omp loop bind(parallel)
|
|
for (int i = 0; i < Length; i++) {
|
|
List[i] = i;
|
|
}
|
|
}
|
|
|
|
inline void run10(int *List, int Length) {
|
|
int i = 0;
|
|
int j = 0;
|
|
#pragma omp teams
|
|
for (int i = 0; i < Length; i++) {
|
|
#pragma omp loop bind(teams)
|
|
for (int j = 0; j < Length; j++) {
|
|
List[i] = i+j;
|
|
}
|
|
}
|
|
}
|
|
|
|
};
|
|
#else
|
|
void test() {
|
|
int List[100];
|
|
|
|
pragma_test pt;
|
|
|
|
pt.run1(List, 100);
|
|
pt.run2(List, 100);
|
|
pt.run3(List, 100);
|
|
pt.run4(List, 100);
|
|
pt.run5(List, 100);
|
|
pt.run6(List, 100);
|
|
pt.run7<2, 4>(List, 100);
|
|
pt.run8(List, 100);
|
|
pt.run9(List, 100);
|
|
pt.run10(List, 100);
|
|
}
|
|
|
|
#endif
|