| 1 |
// { dg-do run } |
| 2 |
|
| 3 |
#include <omp.h> |
| 4 |
typedef __PTRDIFF_TYPE__ ptrdiff_t; |
| 5 |
extern "C" void abort (); |
| 6 |
|
| 7 |
template <typename T> |
| 8 |
class I |
| 9 |
{ |
| 10 |
public: |
| 11 |
typedef ptrdiff_t difference_type; |
| 12 |
I (); |
| 13 |
~I (); |
| 14 |
I (T *); |
| 15 |
I (const I &); |
| 16 |
T &operator * (); |
| 17 |
T *operator -> (); |
| 18 |
T &operator [] (const difference_type &) const; |
| 19 |
I &operator = (const I &); |
| 20 |
I &operator ++ (); |
| 21 |
I operator ++ (int); |
| 22 |
I &operator -- (); |
| 23 |
I operator -- (int); |
| 24 |
I &operator += (const difference_type &); |
| 25 |
I &operator -= (const difference_type &); |
| 26 |
I operator + (const difference_type &) const; |
| 27 |
I operator - (const difference_type &) const; |
| 28 |
template <typename S> friend bool operator == (I<S> &, I<S> &); |
| 29 |
template <typename S> friend bool operator == (const I<S> &, const I<S> &); |
| 30 |
template <typename S> friend bool operator < (I<S> &, I<S> &); |
| 31 |
template <typename S> friend bool operator < (const I<S> &, const I<S> &); |
| 32 |
template <typename S> friend bool operator <= (I<S> &, I<S> &); |
| 33 |
template <typename S> friend bool operator <= (const I<S> &, const I<S> &); |
| 34 |
template <typename S> friend bool operator > (I<S> &, I<S> &); |
| 35 |
template <typename S> friend bool operator > (const I<S> &, const I<S> &); |
| 36 |
template <typename S> friend bool operator >= (I<S> &, I<S> &); |
| 37 |
template <typename S> friend bool operator >= (const I<S> &, const I<S> &); |
| 38 |
template <typename S> friend typename I<S>::difference_type operator - (I<S> &, I<S> &); |
| 39 |
template <typename S> friend typename I<S>::difference_type operator - (const I<S> &, const I<S> &); |
| 40 |
template <typename S> friend I<S> operator + (typename I<S>::difference_type , const I<S> &); |
| 41 |
private: |
| 42 |
T *p; |
| 43 |
}; |
| 44 |
template <typename T> I<T>::I () : p (0) {} |
| 45 |
template <typename T> I<T>::~I () { p = (T *) 0; } |
| 46 |
template <typename T> I<T>::I (T *x) : p (x) {} |
| 47 |
template <typename T> I<T>::I (const I &x) : p (x.p) {} |
| 48 |
template <typename T> T &I<T>::operator * () { return *p; } |
| 49 |
template <typename T> T *I<T>::operator -> () { return p; } |
| 50 |
template <typename T> T &I<T>::operator [] (const difference_type &x) const { return p[x]; } |
| 51 |
template <typename T> I<T> &I<T>::operator = (const I &x) { p = x.p; return *this; } |
| 52 |
template <typename T> I<T> &I<T>::operator ++ () { ++p; return *this; } |
| 53 |
template <typename T> I<T> I<T>::operator ++ (int) { return I (p++); } |
| 54 |
template <typename T> I<T> &I<T>::operator -- () { --p; return *this; } |
| 55 |
template <typename T> I<T> I<T>::operator -- (int) { return I (p--); } |
| 56 |
template <typename T> I<T> &I<T>::operator += (const difference_type &x) { p += x; return *this; } |
| 57 |
template <typename T> I<T> &I<T>::operator -= (const difference_type &x) { p -= x; return *this; } |
| 58 |
template <typename T> I<T> I<T>::operator + (const difference_type &x) const { return I (p + x); } |
| 59 |
template <typename T> I<T> I<T>::operator - (const difference_type &x) const { return I (p - x); } |
| 60 |
template <typename T> bool operator == (I<T> &x, I<T> &y) { return x.p == y.p; } |
| 61 |
template <typename T> bool operator == (const I<T> &x, const I<T> &y) { return x.p == y.p; } |
| 62 |
template <typename T> bool operator != (I<T> &x, I<T> &y) { return !(x == y); } |
| 63 |
template <typename T> bool operator != (const I<T> &x, const I<T> &y) { return !(x == y); } |
| 64 |
template <typename T> bool operator < (I<T> &x, I<T> &y) { return x.p < y.p; } |
| 65 |
template <typename T> bool operator < (const I<T> &x, const I<T> &y) { return x.p < y.p; } |
| 66 |
template <typename T> bool operator <= (I<T> &x, I<T> &y) { return x.p <= y.p; } |
| 67 |
template <typename T> bool operator <= (const I<T> &x, const I<T> &y) { return x.p <= y.p; } |
| 68 |
template <typename T> bool operator > (I<T> &x, I<T> &y) { return x.p > y.p; } |
| 69 |
template <typename T> bool operator > (const I<T> &x, const I<T> &y) { return x.p > y.p; } |
| 70 |
template <typename T> bool operator >= (I<T> &x, I<T> &y) { return x.p >= y.p; } |
| 71 |
template <typename T> bool operator >= (const I<T> &x, const I<T> &y) { return x.p >= y.p; } |
| 72 |
template <typename T> typename I<T>::difference_type operator - (I<T> &x, I<T> &y) { return x.p - y.p; } |
| 73 |
template <typename T> typename I<T>::difference_type operator - (const I<T> &x, const I<T> &y) { return x.p - y.p; } |
| 74 |
template <typename T> I<T> operator + (typename I<T>::difference_type x, const I<T> &y) { return I<T> (x + y.p); } |
| 75 |
|
| 76 |
template <typename T> |
| 77 |
class J |
| 78 |
{ |
| 79 |
public: |
| 80 |
J(const I<T> &x, const I<T> &y) : b (x), e (y) {} |
| 81 |
const I<T> &begin (); |
| 82 |
const I<T> &end (); |
| 83 |
private: |
| 84 |
I<T> b, e; |
| 85 |
}; |
| 86 |
|
| 87 |
template <typename T> const I<T> &J<T>::begin () { return b; } |
| 88 |
template <typename T> const I<T> &J<T>::end () { return e; } |
| 89 |
|
| 90 |
int results[2000]; |
| 91 |
|
| 92 |
void |
| 93 |
f1 (J<int> x, J<int> y, J<int> z) |
| 94 |
{ |
| 95 |
I<int> i, j, k; |
| 96 |
int l, f = 0, n = 0, m = 0; |
| 97 |
#pragma omp parallel shared (i, j, k, l) firstprivate (f) \ |
| 98 |
reduction (+:n, m) num_threads (8) |
| 99 |
{ |
| 100 |
#pragma omp for lastprivate (i, j, k, l) schedule (static, 9) \ |
| 101 |
collapse (4) |
| 102 |
for (i = x.begin (); i < x.end (); ++i) |
| 103 |
for (j = y.begin (); j <= y.end (); j += 1) |
| 104 |
for (l = 0; l < 1; l++) |
| 105 |
for (k = z.begin () + 3; k < z.end () - 3; k++) |
| 106 |
if (omp_get_num_threads () == 8 |
| 107 |
&& ((*i + 2) * 12 + (*j + 5) * 4 + (*k - 13) |
| 108 |
!= (omp_get_thread_num () * 9 + f++))) |
| 109 |
n++; |
| 110 |
else |
| 111 |
m++; |
| 112 |
} |
| 113 |
if (n || i != x.end () || j != y.end () + 1 || k != z.end () - 3 |
| 114 |
|| m != 72 || l != 1) |
| 115 |
abort (); |
| 116 |
} |
| 117 |
|
| 118 |
void |
| 119 |
f2 (J<int> x, J<int> y, J<int> z) |
| 120 |
{ |
| 121 |
int f = 0, n = 0, m = 0; |
| 122 |
#pragma omp parallel for firstprivate (f) reduction (+:n, m) \ |
| 123 |
num_threads (8) schedule (static, 9) \ |
| 124 |
collapse (6 - 2) |
| 125 |
for (I<int> i = x.end () - 1; i >= x.begin (); --i) |
| 126 |
for (int l = -131; l >= -131; l--) |
| 127 |
for (I<int> j = y.end (); j > y.begin () - 1; j -= 1) |
| 128 |
{ |
| 129 |
for (I<int> k = z.end () - 4; k >= z.begin () + 3; k--) |
| 130 |
if (omp_get_num_threads () == 8 |
| 131 |
&& ((3 - *i) * 12 + (-3 - *j) * 4 + (16 - *k) |
| 132 |
!= (omp_get_thread_num () * 9 + f++))) |
| 133 |
n++; |
| 134 |
else |
| 135 |
m++; |
| 136 |
} |
| 137 |
if (n || m != 72) |
| 138 |
abort (); |
| 139 |
} |
| 140 |
|
| 141 |
template <typename T> |
| 142 |
void |
| 143 |
f3 (J<int> x, J<int> y, J<int> z) |
| 144 |
{ |
| 145 |
I<int> i, j, k; |
| 146 |
int l, f = 0, n = 0, m = 0; |
| 147 |
#pragma omp parallel shared (i, j, k, l) firstprivate (f) \ |
| 148 |
reduction (+:n, m) num_threads (8) |
| 149 |
{ |
| 150 |
#pragma omp for lastprivate (i, j, k, l) schedule (static, 9) \ |
| 151 |
collapse (4) |
| 152 |
for (i = x.begin (); i < x.end (); ++i) |
| 153 |
for (j = y.begin (); j <= y.end (); j += 1) |
| 154 |
for (k = z.begin () + 3; k < z.end () - 3; k++) |
| 155 |
for (l = 7; l <= 7; l++) |
| 156 |
if (omp_get_num_threads () == 8 |
| 157 |
&& ((*i + 2) * 12 + (*j + 5) * 4 + (*k - 13) |
| 158 |
!= (omp_get_thread_num () * 9 + f++))) |
| 159 |
n++; |
| 160 |
else |
| 161 |
m++; |
| 162 |
} |
| 163 |
if (n || i != x.end () || j != y.end () + 1 || k != z.end () - 3 |
| 164 |
|| m != 72 || l != 8) |
| 165 |
abort (); |
| 166 |
} |
| 167 |
|
| 168 |
template <typename T> |
| 169 |
void |
| 170 |
f4 (J<int> x, J<int> y, J<int> z) |
| 171 |
{ |
| 172 |
int f = 0, n = 0, m = 0; |
| 173 |
#pragma omp parallel for firstprivate (f) reduction (+:n, m) \ |
| 174 |
num_threads (8) schedule (static, 9) \ |
| 175 |
collapse (5 - 2) |
| 176 |
for (I<int> i = x.end () - 1; i >= x.begin (); --i) |
| 177 |
{ |
| 178 |
for (I<int> j = y.end (); j > y.begin () - 1; j -= 1) |
| 179 |
{ |
| 180 |
for (I<int> k = z.end () - 4; k >= z.begin () + 3; k--) |
| 181 |
if (omp_get_num_threads () == 8 |
| 182 |
&& ((3 - *i) * 12 + (-3 - *j) * 4 + (16 - *k) |
| 183 |
!= (omp_get_thread_num () * 9 + f++))) |
| 184 |
n++; |
| 185 |
else |
| 186 |
m++; |
| 187 |
} |
| 188 |
} |
| 189 |
if (n || m != 72) |
| 190 |
abort (); |
| 191 |
} |
| 192 |
|
| 193 |
template <typename T> |
| 194 |
void |
| 195 |
f5 (J<int> x, J<int> y, J<int> z) |
| 196 |
{ |
| 197 |
I<int> i, j, k; |
| 198 |
int f = 0, n = 0, m = 0; |
| 199 |
#pragma omp parallel shared (i, j, k) firstprivate (f) \ |
| 200 |
reduction (+:n, m) num_threads (8) |
| 201 |
{ |
| 202 |
#pragma omp for lastprivate (i, j, k) schedule (static, 9) \ |
| 203 |
collapse (3) |
| 204 |
for (i = x.begin (); i < x.end (); ++i) |
| 205 |
for (j = y.begin (); j <= y.end (); j += (T) 1) |
| 206 |
{ |
| 207 |
for (k = z.begin () + 3; k < z.end () - 3; k++) |
| 208 |
if (omp_get_num_threads () == 8 |
| 209 |
&& ((*i + 2) * 12 + (*j + 5) * 4 + (*k - 13) |
| 210 |
!= (omp_get_thread_num () * 9 + f++))) |
| 211 |
n++; |
| 212 |
else |
| 213 |
m++; |
| 214 |
} |
| 215 |
} |
| 216 |
if (n || i != x.end () || j != y.end () + 1 || k != z.end () - 3 |
| 217 |
|| m != 72) |
| 218 |
abort (); |
| 219 |
} |
| 220 |
|
| 221 |
template <typename T> |
| 222 |
void |
| 223 |
f6 (J<int> x, J<int> y, J<int> z) |
| 224 |
{ |
| 225 |
int f = 0, n = 0, m = 0; |
| 226 |
#pragma omp parallel for firstprivate (f) reduction (+:n, m) \ |
| 227 |
num_threads (8) schedule (static, 9) \ |
| 228 |
collapse (5 - 2) |
| 229 |
for (I<int> i = x.end () - 1; i >= x.begin (); --i) |
| 230 |
{ |
| 231 |
for (I<int> j = y.end (); j > y.begin () - 1; j -= 1) |
| 232 |
{ |
| 233 |
for (I<int> k = z.end () - 4; k >= z.begin () + (T) 3; k--) |
| 234 |
if (omp_get_num_threads () == 8 |
| 235 |
&& ((3 - *i) * 12 + (-3 - *j) * 4 + (16 - *k) |
| 236 |
!= (omp_get_thread_num () * 9 + f++))) |
| 237 |
n++; |
| 238 |
else |
| 239 |
m++; |
| 240 |
} |
| 241 |
} |
| 242 |
if (n || m != 72) |
| 243 |
abort (); |
| 244 |
} |
| 245 |
|
| 246 |
template <typename T> |
| 247 |
void |
| 248 |
f7 (J<T> x, J<T> y, J<T> z) |
| 249 |
{ |
| 250 |
I<T> i, j, k, o = y.begin (); |
| 251 |
T l, f = 0, n = 0, m = 0; |
| 252 |
#pragma omp parallel shared (i, j, k, l) firstprivate (f) \ |
| 253 |
reduction (+:n, m) num_threads (8) |
| 254 |
{ |
| 255 |
#pragma omp for lastprivate (i, j, k, l) schedule (static, 9) \ |
| 256 |
collapse (4) |
| 257 |
for (i = x.begin (); i < x.end (); ++i) |
| 258 |
for (j = y.begin (); j <= y.end (); j += 1) |
| 259 |
for (l = *o; l <= *o; l = 1 + l) |
| 260 |
for (k = z.begin () + 3; k < z.end () - 3; k++) |
| 261 |
if (omp_get_num_threads () == 8 |
| 262 |
&& ((*i + 2) * 12 + (*j + 5) * 4 + (*k - 13) |
| 263 |
!= (omp_get_thread_num () * 9 + f++))) |
| 264 |
n++; |
| 265 |
else |
| 266 |
m++; |
| 267 |
} |
| 268 |
if (n || i != x.end () || j != y.end () + 1 || k != z.end () - 3 |
| 269 |
|| m != 72 || l != *o + 1) |
| 270 |
abort (); |
| 271 |
} |
| 272 |
|
| 273 |
template <typename T> |
| 274 |
void |
| 275 |
f8 (J<T> x, J<T> y, J<T> z) |
| 276 |
{ |
| 277 |
T f = 0, n = 0, m = 0; |
| 278 |
#pragma omp parallel for firstprivate (f) reduction (+:n, m) \ |
| 279 |
num_threads (8) schedule (static, 9) \ |
| 280 |
collapse (6 - 2) |
| 281 |
for (I<T> i = x.end () - 1; i >= x.begin (); --i) |
| 282 |
for (T l = 0; l < 1; l++) |
| 283 |
for (I<T> j = y.end (); j > y.begin () - 1; j -= 1) |
| 284 |
{ |
| 285 |
for (I<T> k = z.end () - 4; k >= z.begin () + 3; k--) |
| 286 |
if (omp_get_num_threads () == 8 |
| 287 |
&& ((3 - *i) * 12 + (-3 - *j) * 4 + (16 - *k) |
| 288 |
!= (omp_get_thread_num () * 9 + f++))) |
| 289 |
n++; |
| 290 |
else |
| 291 |
m++; |
| 292 |
} |
| 293 |
if (n || m != 72) |
| 294 |
abort (); |
| 295 |
} |
| 296 |
|
| 297 |
template <typename S, typename T> |
| 298 |
void |
| 299 |
f9 (J<T> x, J<T> y, J<T> z) |
| 300 |
{ |
| 301 |
S i, j, k, o = y.begin (); |
| 302 |
T l, f = 0, n = 0, m = 0; |
| 303 |
#pragma omp parallel shared (i, j, k, l) firstprivate (f) \ |
| 304 |
reduction (+:n, m) num_threads (8) |
| 305 |
{ |
| 306 |
#pragma omp for lastprivate (i, j, k, l) schedule (static, 9) \ |
| 307 |
collapse (4) |
| 308 |
for (i = x.begin (); i < x.end (); ++i) |
| 309 |
for (j = y.begin (); j <= y.end (); j += 1) |
| 310 |
for (l = *o; l <= *o; l = 1 + l) |
| 311 |
for (k = z.begin () + 3; k < z.end () - 3; k++) |
| 312 |
if (omp_get_num_threads () == 8 |
| 313 |
&& ((*i + 2) * 12 + (*j + 5) * 4 + (*k - 13) |
| 314 |
!= (omp_get_thread_num () * 9 + f++))) |
| 315 |
n++; |
| 316 |
else |
| 317 |
m++; |
| 318 |
} |
| 319 |
if (n || i != x.end () || j != y.end () + 1 || k != z.end () - 3 |
| 320 |
|| m != 72 || l != *o + 1) |
| 321 |
abort (); |
| 322 |
} |
| 323 |
|
| 324 |
template <typename S, typename T> |
| 325 |
void |
| 326 |
f10 (J<T> x, J<T> y, J<T> z) |
| 327 |
{ |
| 328 |
T f = 0, n = 0, m = 0; |
| 329 |
#pragma omp parallel for firstprivate (f) reduction (+:n, m) \ |
| 330 |
num_threads (8) schedule (static, 9) \ |
| 331 |
collapse (6 - 2) |
| 332 |
for (S i = x.end () - 1; i >= x.begin (); --i) |
| 333 |
for (T l = 0; l < 1; l++) |
| 334 |
for (S j = y.end (); j > y.begin () - 1; j -= 1) |
| 335 |
{ |
| 336 |
for (S k = z.end () - 4; k >= z.begin () + 3; k--) |
| 337 |
if (omp_get_num_threads () == 8 |
| 338 |
&& ((3 - *i) * 12 + (-3 - *j) * 4 + (16 - *k) |
| 339 |
!= (omp_get_thread_num () * 9 + f++))) |
| 340 |
n++; |
| 341 |
else |
| 342 |
m++; |
| 343 |
} |
| 344 |
if (n || m != 72) |
| 345 |
abort (); |
| 346 |
} |
| 347 |
|
| 348 |
int |
| 349 |
main () |
| 350 |
{ |
| 351 |
int a[2000]; |
| 352 |
long b[2000]; |
| 353 |
for (int i = 0; i < 2000; i++) |
| 354 |
{ |
| 355 |
a[i] = i - 1000; |
| 356 |
b[i] = i - 1000; |
| 357 |
} |
| 358 |
J<int> x (&a[998], &a[1004]); |
| 359 |
J<int> y (&a[995], &a[997]); |
| 360 |
J<int> z (&a[1010], &a[1020]); |
| 361 |
f1 (x, y, z); |
| 362 |
f2 (x, y, z); |
| 363 |
f3 <int> (x, y, z); |
| 364 |
f4 <int> (x, y, z); |
| 365 |
f5 <int> (x, y, z); |
| 366 |
f6 <int> (x, y, z); |
| 367 |
f7 <int> (x, y, z); |
| 368 |
f8 <int> (x, y, z); |
| 369 |
f9 <I<int>, int> (x, y, z); |
| 370 |
f10 <I<int>, int> (x, y, z); |
| 371 |
} |