libgomp.info revision 1.12
1 This is libgomp.info, produced by makeinfo version 6.5 from 2 libgomp.texi. 3 4 Copyright (C) 2006-2019 Free Software Foundation, Inc. 5 6 Permission is granted to copy, distribute and/or modify this document 7 under the terms of the GNU Free Documentation License, Version 1.3 or 8 any later version published by the Free Software Foundation; with the 9 Invariant Sections being "Funding Free Software", the Front-Cover texts 10 being (a) (see below), and with the Back-Cover Texts being (b) (see 11 below). A copy of the license is included in the section entitled "GNU 12 Free Documentation License". 13 14 (a) The FSF's Front-Cover Text is: 15 16 A GNU Manual 17 18 (b) The FSF's Back-Cover Text is: 19 20 You have freedom to copy and modify this GNU Manual, like GNU 21 software. Copies published by the Free Software Foundation raise funds 22 for GNU development. 23 INFO-DIR-SECTION GNU Libraries 24 START-INFO-DIR-ENTRY 25 * libgomp: (libgomp). GNU Offloading and Multi Processing Runtime Library. 26 END-INFO-DIR-ENTRY 27 28 This manual documents libgomp, the GNU Offloading and Multi 29 Processing Runtime library. This is the GNU implementation of the 30 OpenMP and OpenACC APIs for parallel and accelerator programming in 31 C/C++ and Fortran. 32 33 Published by the Free Software Foundation 51 Franklin Street, Fifth 34 Floor Boston, MA 02110-1301 USA 35 36 Copyright (C) 2006-2019 Free Software Foundation, Inc. 37 38 Permission is granted to copy, distribute and/or modify this document 39 under the terms of the GNU Free Documentation License, Version 1.3 or 40 any later version published by the Free Software Foundation; with the 41 Invariant Sections being "Funding Free Software", the Front-Cover texts 42 being (a) (see below), and with the Back-Cover Texts being (b) (see 43 below). A copy of the license is included in the section entitled "GNU 44 Free Documentation License". 45 46 (a) The FSF's Front-Cover Text is: 47 48 A GNU Manual 49 50 (b) The FSF's Back-Cover Text is: 51 52 You have freedom to copy and modify this GNU Manual, like GNU 53 software. Copies published by the Free Software Foundation raise funds 54 for GNU development. 55 56 57 File: libgomp.info, Node: Top, Next: Enabling OpenMP, Up: (dir) 58 59 Introduction 60 ************ 61 62 This manual documents the usage of libgomp, the GNU Offloading and Multi 63 Processing Runtime Library. This includes the GNU implementation of the 64 OpenMP (https://www.openmp.org) Application Programming Interface (API) 65 for multi-platform shared-memory parallel programming in C/C++ and 66 Fortran, and the GNU implementation of the OpenACC 67 (https://www.openacc.org) Application Programming Interface (API) for 68 offloading of code to accelerator devices in C/C++ and Fortran. 69 70 Originally, libgomp implemented the GNU OpenMP Runtime Library. 71 Based on this, support for OpenACC and offloading (both OpenACC and 72 OpenMP 4's target construct) has been added later on, and the library's 73 name changed to GNU Offloading and Multi Processing Runtime Library. 74 75 * Menu: 76 77 * Enabling OpenMP:: How to enable OpenMP for your applications. 78 * OpenMP Runtime Library Routines: Runtime Library Routines. 79 The OpenMP runtime application programming 80 interface. 81 * OpenMP Environment Variables: Environment Variables. 82 Influencing OpenMP runtime behavior with 83 environment variables. 84 * Enabling OpenACC:: How to enable OpenACC for your 85 applications. 86 * OpenACC Runtime Library Routines:: The OpenACC runtime application 87 programming interface. 88 * OpenACC Environment Variables:: Influencing OpenACC runtime behavior with 89 environment variables. 90 * CUDA Streams Usage:: Notes on the implementation of 91 asynchronous operations. 92 * OpenACC Library Interoperability:: OpenACC library interoperability with the 93 NVIDIA CUBLAS library. 94 * The libgomp ABI:: Notes on the external ABI presented by libgomp. 95 * Reporting Bugs:: How to report bugs in the GNU Offloading and 96 Multi Processing Runtime Library. 97 * Copying:: GNU general public license says 98 how you can copy and share libgomp. 99 * GNU Free Documentation License:: 100 How you can copy and share this manual. 101 * Funding:: How to help assure continued work for free 102 software. 103 * Library Index:: Index of this documentation. 104 105 106 File: libgomp.info, Node: Enabling OpenMP, Next: Runtime Library Routines, Prev: Top, Up: Top 107 108 1 Enabling OpenMP 109 ***************** 110 111 To activate the OpenMP extensions for C/C++ and Fortran, the 112 compile-time flag '-fopenmp' must be specified. This enables the OpenMP 113 directive '#pragma omp' in C/C++ and '!$omp' directives in free form, 114 'c$omp', '*$omp' and '!$omp' directives in fixed form, '!$' conditional 115 compilation sentinels in free form and 'c$', '*$' and '!$' sentinels in 116 fixed form, for Fortran. The flag also arranges for automatic linking 117 of the OpenMP runtime library (*note Runtime Library Routines::). 118 119 A complete description of all OpenMP directives accepted may be found 120 in the OpenMP Application Program Interface (https://www.openmp.org) 121 manual, version 4.5. 122 123 124 File: libgomp.info, Node: Runtime Library Routines, Next: Environment Variables, Prev: Enabling OpenMP, Up: Top 125 126 2 OpenMP Runtime Library Routines 127 ********************************* 128 129 The runtime routines described here are defined by Section 3 of the 130 OpenMP specification in version 4.5. The routines are structured in 131 following three parts: 132 133 * Menu: 134 135 Control threads, processors and the parallel environment. They have C 136 linkage, and do not throw exceptions. 137 138 * omp_get_active_level:: Number of active parallel regions 139 * omp_get_ancestor_thread_num:: Ancestor thread ID 140 * omp_get_cancellation:: Whether cancellation support is enabled 141 * omp_get_default_device:: Get the default device for target regions 142 * omp_get_dynamic:: Dynamic teams setting 143 * omp_get_level:: Number of parallel regions 144 * omp_get_max_active_levels:: Maximum number of active regions 145 * omp_get_max_task_priority:: Maximum task priority value that can be set 146 * omp_get_max_threads:: Maximum number of threads of parallel region 147 * omp_get_nested:: Nested parallel regions 148 * omp_get_num_devices:: Number of target devices 149 * omp_get_num_procs:: Number of processors online 150 * omp_get_num_teams:: Number of teams 151 * omp_get_num_threads:: Size of the active team 152 * omp_get_proc_bind:: Whether theads may be moved between CPUs 153 * omp_get_schedule:: Obtain the runtime scheduling method 154 * omp_get_team_num:: Get team number 155 * omp_get_team_size:: Number of threads in a team 156 * omp_get_thread_limit:: Maximum number of threads 157 * omp_get_thread_num:: Current thread ID 158 * omp_in_parallel:: Whether a parallel region is active 159 * omp_in_final:: Whether in final or included task region 160 * omp_is_initial_device:: Whether executing on the host device 161 * omp_set_default_device:: Set the default device for target regions 162 * omp_set_dynamic:: Enable/disable dynamic teams 163 * omp_set_max_active_levels:: Limits the number of active parallel regions 164 * omp_set_nested:: Enable/disable nested parallel regions 165 * omp_set_num_threads:: Set upper team size limit 166 * omp_set_schedule:: Set the runtime scheduling method 167 168 Initialize, set, test, unset and destroy simple and nested locks. 169 170 * omp_init_lock:: Initialize simple lock 171 * omp_set_lock:: Wait for and set simple lock 172 * omp_test_lock:: Test and set simple lock if available 173 * omp_unset_lock:: Unset simple lock 174 * omp_destroy_lock:: Destroy simple lock 175 * omp_init_nest_lock:: Initialize nested lock 176 * omp_set_nest_lock:: Wait for and set simple lock 177 * omp_test_nest_lock:: Test and set nested lock if available 178 * omp_unset_nest_lock:: Unset nested lock 179 * omp_destroy_nest_lock:: Destroy nested lock 180 181 Portable, thread-based, wall clock timer. 182 183 * omp_get_wtick:: Get timer precision. 184 * omp_get_wtime:: Elapsed wall clock time. 185 186 187 File: libgomp.info, Node: omp_get_active_level, Next: omp_get_ancestor_thread_num, Up: Runtime Library Routines 188 189 2.1 'omp_get_active_level' - Number of parallel regions 190 ======================================================= 191 192 _Description_: 193 This function returns the nesting level for the active parallel 194 blocks, which enclose the calling call. 195 196 _C/C++_ 197 _Prototype_: 'int omp_get_active_level(void);' 198 199 _Fortran_: 200 _Interface_: 'integer function omp_get_active_level()' 201 202 _See also_: 203 *note omp_get_level::, *note omp_get_max_active_levels::, *note 204 omp_set_max_active_levels:: 205 206 _Reference_: 207 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.20. 208 209 210 File: libgomp.info, Node: omp_get_ancestor_thread_num, Next: omp_get_cancellation, Prev: omp_get_active_level, Up: Runtime Library Routines 211 212 2.2 'omp_get_ancestor_thread_num' - Ancestor thread ID 213 ====================================================== 214 215 _Description_: 216 This function returns the thread identification number for the 217 given nesting level of the current thread. For values of LEVEL 218 outside zero to 'omp_get_level' -1 is returned; if LEVEL is 219 'omp_get_level' the result is identical to 'omp_get_thread_num'. 220 221 _C/C++_ 222 _Prototype_: 'int omp_get_ancestor_thread_num(int level);' 223 224 _Fortran_: 225 _Interface_: 'integer function omp_get_ancestor_thread_num(level)' 226 'integer level' 227 228 _See also_: 229 *note omp_get_level::, *note omp_get_thread_num::, *note 230 omp_get_team_size:: 231 232 _Reference_: 233 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.18. 234 235 236 File: libgomp.info, Node: omp_get_cancellation, Next: omp_get_default_device, Prev: omp_get_ancestor_thread_num, Up: Runtime Library Routines 237 238 2.3 'omp_get_cancellation' - Whether cancellation support is enabled 239 ==================================================================== 240 241 _Description_: 242 This function returns 'true' if cancellation is activated, 'false' 243 otherwise. Here, 'true' and 'false' represent their 244 language-specific counterparts. Unless 'OMP_CANCELLATION' is set 245 true, cancellations are deactivated. 246 247 _C/C++_: 248 _Prototype_: 'int omp_get_cancellation(void);' 249 250 _Fortran_: 251 _Interface_: 'logical function omp_get_cancellation()' 252 253 _See also_: 254 *note OMP_CANCELLATION:: 255 256 _Reference_: 257 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.9. 258 259 260 File: libgomp.info, Node: omp_get_default_device, Next: omp_get_dynamic, Prev: omp_get_cancellation, Up: Runtime Library Routines 261 262 2.4 'omp_get_default_device' - Get the default device for target regions 263 ======================================================================== 264 265 _Description_: 266 Get the default device for target regions without device clause. 267 268 _C/C++_: 269 _Prototype_: 'int omp_get_default_device(void);' 270 271 _Fortran_: 272 _Interface_: 'integer function omp_get_default_device()' 273 274 _See also_: 275 *note OMP_DEFAULT_DEVICE::, *note omp_set_default_device:: 276 277 _Reference_: 278 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.30. 279 280 281 File: libgomp.info, Node: omp_get_dynamic, Next: omp_get_level, Prev: omp_get_default_device, Up: Runtime Library Routines 282 283 2.5 'omp_get_dynamic' - Dynamic teams setting 284 ============================================= 285 286 _Description_: 287 This function returns 'true' if enabled, 'false' otherwise. Here, 288 'true' and 'false' represent their language-specific counterparts. 289 290 The dynamic team setting may be initialized at startup by the 291 'OMP_DYNAMIC' environment variable or at runtime using 292 'omp_set_dynamic'. If undefined, dynamic adjustment is disabled by 293 default. 294 295 _C/C++_: 296 _Prototype_: 'int omp_get_dynamic(void);' 297 298 _Fortran_: 299 _Interface_: 'logical function omp_get_dynamic()' 300 301 _See also_: 302 *note omp_set_dynamic::, *note OMP_DYNAMIC:: 303 304 _Reference_: 305 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.8. 306 307 308 File: libgomp.info, Node: omp_get_level, Next: omp_get_max_active_levels, Prev: omp_get_dynamic, Up: Runtime Library Routines 309 310 2.6 'omp_get_level' - Obtain the current nesting level 311 ====================================================== 312 313 _Description_: 314 This function returns the nesting level for the parallel blocks, 315 which enclose the calling call. 316 317 _C/C++_ 318 _Prototype_: 'int omp_get_level(void);' 319 320 _Fortran_: 321 _Interface_: 'integer function omp_level()' 322 323 _See also_: 324 *note omp_get_active_level:: 325 326 _Reference_: 327 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.17. 328 329 330 File: libgomp.info, Node: omp_get_max_active_levels, Next: omp_get_max_task_priority, Prev: omp_get_level, Up: Runtime Library Routines 331 332 2.7 'omp_get_max_active_levels' - Maximum number of active regions 333 ================================================================== 334 335 _Description_: 336 This function obtains the maximum allowed number of nested, active 337 parallel regions. 338 339 _C/C++_ 340 _Prototype_: 'int omp_get_max_active_levels(void);' 341 342 _Fortran_: 343 _Interface_: 'integer function omp_get_max_active_levels()' 344 345 _See also_: 346 *note omp_set_max_active_levels::, *note omp_get_active_level:: 347 348 _Reference_: 349 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.16. 350 351 352 File: libgomp.info, Node: omp_get_max_task_priority, Next: omp_get_max_threads, Prev: omp_get_max_active_levels, Up: Runtime Library Routines 353 354 2.8 'omp_get_max_task_priority' - Maximum priority value 355 ======================================================== 356 357 that can be set for tasks. 358 _Description_: 359 This function obtains the maximum allowed priority number for 360 tasks. 361 362 _C/C++_ 363 _Prototype_: 'int omp_get_max_task_priority(void);' 364 365 _Fortran_: 366 _Interface_: 'integer function omp_get_max_task_priority()' 367 368 _Reference_: 369 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.29. 370 371 372 File: libgomp.info, Node: omp_get_max_threads, Next: omp_get_nested, Prev: omp_get_max_task_priority, Up: Runtime Library Routines 373 374 2.9 'omp_get_max_threads' - Maximum number of threads of parallel region 375 ======================================================================== 376 377 _Description_: 378 Return the maximum number of threads used for the current parallel 379 region that does not use the clause 'num_threads'. 380 381 _C/C++_: 382 _Prototype_: 'int omp_get_max_threads(void);' 383 384 _Fortran_: 385 _Interface_: 'integer function omp_get_max_threads()' 386 387 _See also_: 388 *note omp_set_num_threads::, *note omp_set_dynamic::, *note 389 omp_get_thread_limit:: 390 391 _Reference_: 392 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.3. 393 394 395 File: libgomp.info, Node: omp_get_nested, Next: omp_get_num_devices, Prev: omp_get_max_threads, Up: Runtime Library Routines 396 397 2.10 'omp_get_nested' - Nested parallel regions 398 =============================================== 399 400 _Description_: 401 This function returns 'true' if nested parallel regions are 402 enabled, 'false' otherwise. Here, 'true' and 'false' represent 403 their language-specific counterparts. 404 405 Nested parallel regions may be initialized at startup by the 406 'OMP_NESTED' environment variable or at runtime using 407 'omp_set_nested'. If undefined, nested parallel regions are 408 disabled by default. 409 410 _C/C++_: 411 _Prototype_: 'int omp_get_nested(void);' 412 413 _Fortran_: 414 _Interface_: 'logical function omp_get_nested()' 415 416 _See also_: 417 *note omp_set_nested::, *note OMP_NESTED:: 418 419 _Reference_: 420 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.11. 421 422 423 File: libgomp.info, Node: omp_get_num_devices, Next: omp_get_num_procs, Prev: omp_get_nested, Up: Runtime Library Routines 424 425 2.11 'omp_get_num_devices' - Number of target devices 426 ===================================================== 427 428 _Description_: 429 Returns the number of target devices. 430 431 _C/C++_: 432 _Prototype_: 'int omp_get_num_devices(void);' 433 434 _Fortran_: 435 _Interface_: 'integer function omp_get_num_devices()' 436 437 _Reference_: 438 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.31. 439 440 441 File: libgomp.info, Node: omp_get_num_procs, Next: omp_get_num_teams, Prev: omp_get_num_devices, Up: Runtime Library Routines 442 443 2.12 'omp_get_num_procs' - Number of processors online 444 ====================================================== 445 446 _Description_: 447 Returns the number of processors online on that device. 448 449 _C/C++_: 450 _Prototype_: 'int omp_get_num_procs(void);' 451 452 _Fortran_: 453 _Interface_: 'integer function omp_get_num_procs()' 454 455 _Reference_: 456 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.5. 457 458 459 File: libgomp.info, Node: omp_get_num_teams, Next: omp_get_num_threads, Prev: omp_get_num_procs, Up: Runtime Library Routines 460 461 2.13 'omp_get_num_teams' - Number of teams 462 ========================================== 463 464 _Description_: 465 Returns the number of teams in the current team region. 466 467 _C/C++_: 468 _Prototype_: 'int omp_get_num_teams(void);' 469 470 _Fortran_: 471 _Interface_: 'integer function omp_get_num_teams()' 472 473 _Reference_: 474 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.32. 475 476 477 File: libgomp.info, Node: omp_get_num_threads, Next: omp_get_proc_bind, Prev: omp_get_num_teams, Up: Runtime Library Routines 478 479 2.14 'omp_get_num_threads' - Size of the active team 480 ==================================================== 481 482 _Description_: 483 Returns the number of threads in the current team. In a sequential 484 section of the program 'omp_get_num_threads' returns 1. 485 486 The default team size may be initialized at startup by the 487 'OMP_NUM_THREADS' environment variable. At runtime, the size of 488 the current team may be set either by the 'NUM_THREADS' clause or 489 by 'omp_set_num_threads'. If none of the above were used to define 490 a specific value and 'OMP_DYNAMIC' is disabled, one thread per CPU 491 online is used. 492 493 _C/C++_: 494 _Prototype_: 'int omp_get_num_threads(void);' 495 496 _Fortran_: 497 _Interface_: 'integer function omp_get_num_threads()' 498 499 _See also_: 500 *note omp_get_max_threads::, *note omp_set_num_threads::, *note 501 OMP_NUM_THREADS:: 502 503 _Reference_: 504 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.2. 505 506 507 File: libgomp.info, Node: omp_get_proc_bind, Next: omp_get_schedule, Prev: omp_get_num_threads, Up: Runtime Library Routines 508 509 2.15 'omp_get_proc_bind' - Whether theads may be moved between CPUs 510 =================================================================== 511 512 _Description_: 513 This functions returns the currently active thread affinity policy, 514 which is set via 'OMP_PROC_BIND'. Possible values are 515 'omp_proc_bind_false', 'omp_proc_bind_true', 516 'omp_proc_bind_master', 'omp_proc_bind_close' and 517 'omp_proc_bind_spread'. 518 519 _C/C++_: 520 _Prototype_: 'omp_proc_bind_t omp_get_proc_bind(void);' 521 522 _Fortran_: 523 _Interface_: 'integer(kind=omp_proc_bind_kind) function 524 omp_get_proc_bind()' 525 526 _See also_: 527 *note OMP_PROC_BIND::, *note OMP_PLACES::, *note 528 GOMP_CPU_AFFINITY::, 529 530 _Reference_: 531 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.22. 532 533 534 File: libgomp.info, Node: omp_get_schedule, Next: omp_get_team_num, Prev: omp_get_proc_bind, Up: Runtime Library Routines 535 536 2.16 'omp_get_schedule' - Obtain the runtime scheduling method 537 ============================================================== 538 539 _Description_: 540 Obtain the runtime scheduling method. The KIND argument will be 541 set to the value 'omp_sched_static', 'omp_sched_dynamic', 542 'omp_sched_guided' or 'omp_sched_auto'. The second argument, 543 CHUNK_SIZE, is set to the chunk size. 544 545 _C/C++_ 546 _Prototype_: 'void omp_get_schedule(omp_sched_t *kind, int 547 *chunk_size);' 548 549 _Fortran_: 550 _Interface_: 'subroutine omp_get_schedule(kind, chunk_size)' 551 'integer(kind=omp_sched_kind) kind' 552 'integer chunk_size' 553 554 _See also_: 555 *note omp_set_schedule::, *note OMP_SCHEDULE:: 556 557 _Reference_: 558 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.13. 559 560 561 File: libgomp.info, Node: omp_get_team_num, Next: omp_get_team_size, Prev: omp_get_schedule, Up: Runtime Library Routines 562 563 2.17 'omp_get_team_num' - Get team number 564 ========================================= 565 566 _Description_: 567 Returns the team number of the calling thread. 568 569 _C/C++_: 570 _Prototype_: 'int omp_get_team_num(void);' 571 572 _Fortran_: 573 _Interface_: 'integer function omp_get_team_num()' 574 575 _Reference_: 576 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.33. 577 578 579 File: libgomp.info, Node: omp_get_team_size, Next: omp_get_thread_limit, Prev: omp_get_team_num, Up: Runtime Library Routines 580 581 2.18 'omp_get_team_size' - Number of threads in a team 582 ====================================================== 583 584 _Description_: 585 This function returns the number of threads in a thread team to 586 which either the current thread or its ancestor belongs. For 587 values of LEVEL outside zero to 'omp_get_level', -1 is returned; if 588 LEVEL is zero, 1 is returned, and for 'omp_get_level', the result 589 is identical to 'omp_get_num_threads'. 590 591 _C/C++_: 592 _Prototype_: 'int omp_get_team_size(int level);' 593 594 _Fortran_: 595 _Interface_: 'integer function omp_get_team_size(level)' 596 'integer level' 597 598 _See also_: 599 *note omp_get_num_threads::, *note omp_get_level::, *note 600 omp_get_ancestor_thread_num:: 601 602 _Reference_: 603 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.19. 604 605 606 File: libgomp.info, Node: omp_get_thread_limit, Next: omp_get_thread_num, Prev: omp_get_team_size, Up: Runtime Library Routines 607 608 2.19 'omp_get_thread_limit' - Maximum number of threads 609 ======================================================= 610 611 _Description_: 612 Return the maximum number of threads of the program. 613 614 _C/C++_: 615 _Prototype_: 'int omp_get_thread_limit(void);' 616 617 _Fortran_: 618 _Interface_: 'integer function omp_get_thread_limit()' 619 620 _See also_: 621 *note omp_get_max_threads::, *note OMP_THREAD_LIMIT:: 622 623 _Reference_: 624 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.14. 625 626 627 File: libgomp.info, Node: omp_get_thread_num, Next: omp_in_parallel, Prev: omp_get_thread_limit, Up: Runtime Library Routines 628 629 2.20 'omp_get_thread_num' - Current thread ID 630 ============================================= 631 632 _Description_: 633 Returns a unique thread identification number within the current 634 team. In a sequential parts of the program, 'omp_get_thread_num' 635 always returns 0. In parallel regions the return value varies from 636 0 to 'omp_get_num_threads'-1 inclusive. The return value of the 637 master thread of a team is always 0. 638 639 _C/C++_: 640 _Prototype_: 'int omp_get_thread_num(void);' 641 642 _Fortran_: 643 _Interface_: 'integer function omp_get_thread_num()' 644 645 _See also_: 646 *note omp_get_num_threads::, *note omp_get_ancestor_thread_num:: 647 648 _Reference_: 649 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.4. 650 651 652 File: libgomp.info, Node: omp_in_parallel, Next: omp_in_final, Prev: omp_get_thread_num, Up: Runtime Library Routines 653 654 2.21 'omp_in_parallel' - Whether a parallel region is active 655 ============================================================ 656 657 _Description_: 658 This function returns 'true' if currently running in parallel, 659 'false' otherwise. Here, 'true' and 'false' represent their 660 language-specific counterparts. 661 662 _C/C++_: 663 _Prototype_: 'int omp_in_parallel(void);' 664 665 _Fortran_: 666 _Interface_: 'logical function omp_in_parallel()' 667 668 _Reference_: 669 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.6. 670 671 672 File: libgomp.info, Node: omp_in_final, Next: omp_is_initial_device, Prev: omp_in_parallel, Up: Runtime Library Routines 673 674 2.22 'omp_in_final' - Whether in final or included task region 675 ============================================================== 676 677 _Description_: 678 This function returns 'true' if currently running in a final or 679 included task region, 'false' otherwise. Here, 'true' and 'false' 680 represent their language-specific counterparts. 681 682 _C/C++_: 683 _Prototype_: 'int omp_in_final(void);' 684 685 _Fortran_: 686 _Interface_: 'logical function omp_in_final()' 687 688 _Reference_: 689 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.21. 690 691 692 File: libgomp.info, Node: omp_is_initial_device, Next: omp_set_default_device, Prev: omp_in_final, Up: Runtime Library Routines 693 694 2.23 'omp_is_initial_device' - Whether executing on the host device 695 =================================================================== 696 697 _Description_: 698 This function returns 'true' if currently running on the host 699 device, 'false' otherwise. Here, 'true' and 'false' represent 700 their language-specific counterparts. 701 702 _C/C++_: 703 _Prototype_: 'int omp_is_initial_device(void);' 704 705 _Fortran_: 706 _Interface_: 'logical function omp_is_initial_device()' 707 708 _Reference_: 709 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.34. 710 711 712 File: libgomp.info, Node: omp_set_default_device, Next: omp_set_dynamic, Prev: omp_is_initial_device, Up: Runtime Library Routines 713 714 2.24 'omp_set_default_device' - Set the default device for target regions 715 ========================================================================= 716 717 _Description_: 718 Set the default device for target regions without device clause. 719 The argument shall be a nonnegative device number. 720 721 _C/C++_: 722 _Prototype_: 'void omp_set_default_device(int device_num);' 723 724 _Fortran_: 725 _Interface_: 'subroutine omp_set_default_device(device_num)' 726 'integer device_num' 727 728 _See also_: 729 *note OMP_DEFAULT_DEVICE::, *note omp_get_default_device:: 730 731 _Reference_: 732 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.29. 733 734 735 File: libgomp.info, Node: omp_set_dynamic, Next: omp_set_max_active_levels, Prev: omp_set_default_device, Up: Runtime Library Routines 736 737 2.25 'omp_set_dynamic' - Enable/disable dynamic teams 738 ===================================================== 739 740 _Description_: 741 Enable or disable the dynamic adjustment of the number of threads 742 within a team. The function takes the language-specific equivalent 743 of 'true' and 'false', where 'true' enables dynamic adjustment of 744 team sizes and 'false' disables it. 745 746 _C/C++_: 747 _Prototype_: 'void omp_set_dynamic(int dynamic_threads);' 748 749 _Fortran_: 750 _Interface_: 'subroutine omp_set_dynamic(dynamic_threads)' 751 'logical, intent(in) :: dynamic_threads' 752 753 _See also_: 754 *note OMP_DYNAMIC::, *note omp_get_dynamic:: 755 756 _Reference_: 757 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.7. 758 759 760 File: libgomp.info, Node: omp_set_max_active_levels, Next: omp_set_nested, Prev: omp_set_dynamic, Up: Runtime Library Routines 761 762 2.26 'omp_set_max_active_levels' - Limits the number of active parallel regions 763 =============================================================================== 764 765 _Description_: 766 This function limits the maximum allowed number of nested, active 767 parallel regions. 768 769 _C/C++_ 770 _Prototype_: 'void omp_set_max_active_levels(int max_levels);' 771 772 _Fortran_: 773 _Interface_: 'subroutine omp_set_max_active_levels(max_levels)' 774 'integer max_levels' 775 776 _See also_: 777 *note omp_get_max_active_levels::, *note omp_get_active_level:: 778 779 _Reference_: 780 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.15. 781 782 783 File: libgomp.info, Node: omp_set_nested, Next: omp_set_num_threads, Prev: omp_set_max_active_levels, Up: Runtime Library Routines 784 785 2.27 'omp_set_nested' - Enable/disable nested parallel regions 786 ============================================================== 787 788 _Description_: 789 Enable or disable nested parallel regions, i.e., whether team 790 members are allowed to create new teams. The function takes the 791 language-specific equivalent of 'true' and 'false', where 'true' 792 enables dynamic adjustment of team sizes and 'false' disables it. 793 794 _C/C++_: 795 _Prototype_: 'void omp_set_nested(int nested);' 796 797 _Fortran_: 798 _Interface_: 'subroutine omp_set_nested(nested)' 799 'logical, intent(in) :: nested' 800 801 _See also_: 802 *note OMP_NESTED::, *note omp_get_nested:: 803 804 _Reference_: 805 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.10. 806 807 808 File: libgomp.info, Node: omp_set_num_threads, Next: omp_set_schedule, Prev: omp_set_nested, Up: Runtime Library Routines 809 810 2.28 'omp_set_num_threads' - Set upper team size limit 811 ====================================================== 812 813 _Description_: 814 Specifies the number of threads used by default in subsequent 815 parallel sections, if those do not specify a 'num_threads' clause. 816 The argument of 'omp_set_num_threads' shall be a positive integer. 817 818 _C/C++_: 819 _Prototype_: 'void omp_set_num_threads(int num_threads);' 820 821 _Fortran_: 822 _Interface_: 'subroutine omp_set_num_threads(num_threads)' 823 'integer, intent(in) :: num_threads' 824 825 _See also_: 826 *note OMP_NUM_THREADS::, *note omp_get_num_threads::, *note 827 omp_get_max_threads:: 828 829 _Reference_: 830 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.1. 831 832 833 File: libgomp.info, Node: omp_set_schedule, Next: omp_init_lock, Prev: omp_set_num_threads, Up: Runtime Library Routines 834 835 2.29 'omp_set_schedule' - Set the runtime scheduling method 836 =========================================================== 837 838 _Description_: 839 Sets the runtime scheduling method. The KIND argument can have the 840 value 'omp_sched_static', 'omp_sched_dynamic', 'omp_sched_guided' 841 or 'omp_sched_auto'. Except for 'omp_sched_auto', the chunk size 842 is set to the value of CHUNK_SIZE if positive, or to the default 843 value if zero or negative. For 'omp_sched_auto' the CHUNK_SIZE 844 argument is ignored. 845 846 _C/C++_ 847 _Prototype_: 'void omp_set_schedule(omp_sched_t kind, int 848 chunk_size);' 849 850 _Fortran_: 851 _Interface_: 'subroutine omp_set_schedule(kind, chunk_size)' 852 'integer(kind=omp_sched_kind) kind' 853 'integer chunk_size' 854 855 _See also_: 856 *note omp_get_schedule:: *note OMP_SCHEDULE:: 857 858 _Reference_: 859 OpenMP specification v4.5 (https://www.openmp.org), Section 3.2.12. 860 861 862 File: libgomp.info, Node: omp_init_lock, Next: omp_set_lock, Prev: omp_set_schedule, Up: Runtime Library Routines 863 864 2.30 'omp_init_lock' - Initialize simple lock 865 ============================================= 866 867 _Description_: 868 Initialize a simple lock. After initialization, the lock is in an 869 unlocked state. 870 871 _C/C++_: 872 _Prototype_: 'void omp_init_lock(omp_lock_t *lock);' 873 874 _Fortran_: 875 _Interface_: 'subroutine omp_init_lock(svar)' 876 'integer(omp_lock_kind), intent(out) :: svar' 877 878 _See also_: 879 *note omp_destroy_lock:: 880 881 _Reference_: 882 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.1. 883 884 885 File: libgomp.info, Node: omp_set_lock, Next: omp_test_lock, Prev: omp_init_lock, Up: Runtime Library Routines 886 887 2.31 'omp_set_lock' - Wait for and set simple lock 888 ================================================== 889 890 _Description_: 891 Before setting a simple lock, the lock variable must be initialized 892 by 'omp_init_lock'. The calling thread is blocked until the lock 893 is available. If the lock is already held by the current thread, a 894 deadlock occurs. 895 896 _C/C++_: 897 _Prototype_: 'void omp_set_lock(omp_lock_t *lock);' 898 899 _Fortran_: 900 _Interface_: 'subroutine omp_set_lock(svar)' 901 'integer(omp_lock_kind), intent(inout) :: svar' 902 903 _See also_: 904 *note omp_init_lock::, *note omp_test_lock::, *note 905 omp_unset_lock:: 906 907 _Reference_: 908 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.4. 909 910 911 File: libgomp.info, Node: omp_test_lock, Next: omp_unset_lock, Prev: omp_set_lock, Up: Runtime Library Routines 912 913 2.32 'omp_test_lock' - Test and set simple lock if available 914 ============================================================ 915 916 _Description_: 917 Before setting a simple lock, the lock variable must be initialized 918 by 'omp_init_lock'. Contrary to 'omp_set_lock', 'omp_test_lock' 919 does not block if the lock is not available. This function returns 920 'true' upon success, 'false' otherwise. Here, 'true' and 'false' 921 represent their language-specific counterparts. 922 923 _C/C++_: 924 _Prototype_: 'int omp_test_lock(omp_lock_t *lock);' 925 926 _Fortran_: 927 _Interface_: 'logical function omp_test_lock(svar)' 928 'integer(omp_lock_kind), intent(inout) :: svar' 929 930 _See also_: 931 *note omp_init_lock::, *note omp_set_lock::, *note omp_set_lock:: 932 933 _Reference_: 934 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.6. 935 936 937 File: libgomp.info, Node: omp_unset_lock, Next: omp_destroy_lock, Prev: omp_test_lock, Up: Runtime Library Routines 938 939 2.33 'omp_unset_lock' - Unset simple lock 940 ========================================= 941 942 _Description_: 943 A simple lock about to be unset must have been locked by 944 'omp_set_lock' or 'omp_test_lock' before. In addition, the lock 945 must be held by the thread calling 'omp_unset_lock'. Then, the 946 lock becomes unlocked. If one or more threads attempted to set the 947 lock before, one of them is chosen to, again, set the lock to 948 itself. 949 950 _C/C++_: 951 _Prototype_: 'void omp_unset_lock(omp_lock_t *lock);' 952 953 _Fortran_: 954 _Interface_: 'subroutine omp_unset_lock(svar)' 955 'integer(omp_lock_kind), intent(inout) :: svar' 956 957 _See also_: 958 *note omp_set_lock::, *note omp_test_lock:: 959 960 _Reference_: 961 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.5. 962 963 964 File: libgomp.info, Node: omp_destroy_lock, Next: omp_init_nest_lock, Prev: omp_unset_lock, Up: Runtime Library Routines 965 966 2.34 'omp_destroy_lock' - Destroy simple lock 967 ============================================= 968 969 _Description_: 970 Destroy a simple lock. In order to be destroyed, a simple lock 971 must be in the unlocked state. 972 973 _C/C++_: 974 _Prototype_: 'void omp_destroy_lock(omp_lock_t *lock);' 975 976 _Fortran_: 977 _Interface_: 'subroutine omp_destroy_lock(svar)' 978 'integer(omp_lock_kind), intent(inout) :: svar' 979 980 _See also_: 981 *note omp_init_lock:: 982 983 _Reference_: 984 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.3. 985 986 987 File: libgomp.info, Node: omp_init_nest_lock, Next: omp_set_nest_lock, Prev: omp_destroy_lock, Up: Runtime Library Routines 988 989 2.35 'omp_init_nest_lock' - Initialize nested lock 990 ================================================== 991 992 _Description_: 993 Initialize a nested lock. After initialization, the lock is in an 994 unlocked state and the nesting count is set to zero. 995 996 _C/C++_: 997 _Prototype_: 'void omp_init_nest_lock(omp_nest_lock_t *lock);' 998 999 _Fortran_: 1000 _Interface_: 'subroutine omp_init_nest_lock(nvar)' 1001 'integer(omp_nest_lock_kind), intent(out) :: nvar' 1002 1003 _See also_: 1004 *note omp_destroy_nest_lock:: 1005 1006 _Reference_: 1007 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.1. 1008 1009 1010 File: libgomp.info, Node: omp_set_nest_lock, Next: omp_test_nest_lock, Prev: omp_init_nest_lock, Up: Runtime Library Routines 1011 1012 2.36 'omp_set_nest_lock' - Wait for and set nested lock 1013 ======================================================= 1014 1015 _Description_: 1016 Before setting a nested lock, the lock variable must be initialized 1017 by 'omp_init_nest_lock'. The calling thread is blocked until the 1018 lock is available. If the lock is already held by the current 1019 thread, the nesting count for the lock is incremented. 1020 1021 _C/C++_: 1022 _Prototype_: 'void omp_set_nest_lock(omp_nest_lock_t *lock);' 1023 1024 _Fortran_: 1025 _Interface_: 'subroutine omp_set_nest_lock(nvar)' 1026 'integer(omp_nest_lock_kind), intent(inout) :: nvar' 1027 1028 _See also_: 1029 *note omp_init_nest_lock::, *note omp_unset_nest_lock:: 1030 1031 _Reference_: 1032 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.4. 1033 1034 1035 File: libgomp.info, Node: omp_test_nest_lock, Next: omp_unset_nest_lock, Prev: omp_set_nest_lock, Up: Runtime Library Routines 1036 1037 2.37 'omp_test_nest_lock' - Test and set nested lock if available 1038 ================================================================= 1039 1040 _Description_: 1041 Before setting a nested lock, the lock variable must be initialized 1042 by 'omp_init_nest_lock'. Contrary to 'omp_set_nest_lock', 1043 'omp_test_nest_lock' does not block if the lock is not available. 1044 If the lock is already held by the current thread, the new nesting 1045 count is returned. Otherwise, the return value equals zero. 1046 1047 _C/C++_: 1048 _Prototype_: 'int omp_test_nest_lock(omp_nest_lock_t *lock);' 1049 1050 _Fortran_: 1051 _Interface_: 'logical function omp_test_nest_lock(nvar)' 1052 'integer(omp_nest_lock_kind), intent(inout) :: nvar' 1053 1054 _See also_: 1055 *note omp_init_lock::, *note omp_set_lock::, *note omp_set_lock:: 1056 1057 _Reference_: 1058 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.6. 1059 1060 1061 File: libgomp.info, Node: omp_unset_nest_lock, Next: omp_destroy_nest_lock, Prev: omp_test_nest_lock, Up: Runtime Library Routines 1062 1063 2.38 'omp_unset_nest_lock' - Unset nested lock 1064 ============================================== 1065 1066 _Description_: 1067 A nested lock about to be unset must have been locked by 1068 'omp_set_nested_lock' or 'omp_test_nested_lock' before. In 1069 addition, the lock must be held by the thread calling 1070 'omp_unset_nested_lock'. If the nesting count drops to zero, the 1071 lock becomes unlocked. If one ore more threads attempted to set 1072 the lock before, one of them is chosen to, again, set the lock to 1073 itself. 1074 1075 _C/C++_: 1076 _Prototype_: 'void omp_unset_nest_lock(omp_nest_lock_t *lock);' 1077 1078 _Fortran_: 1079 _Interface_: 'subroutine omp_unset_nest_lock(nvar)' 1080 'integer(omp_nest_lock_kind), intent(inout) :: nvar' 1081 1082 _See also_: 1083 *note omp_set_nest_lock:: 1084 1085 _Reference_: 1086 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.5. 1087 1088 1089 File: libgomp.info, Node: omp_destroy_nest_lock, Next: omp_get_wtick, Prev: omp_unset_nest_lock, Up: Runtime Library Routines 1090 1091 2.39 'omp_destroy_nest_lock' - Destroy nested lock 1092 ================================================== 1093 1094 _Description_: 1095 Destroy a nested lock. In order to be destroyed, a nested lock 1096 must be in the unlocked state and its nesting count must equal 1097 zero. 1098 1099 _C/C++_: 1100 _Prototype_: 'void omp_destroy_nest_lock(omp_nest_lock_t *);' 1101 1102 _Fortran_: 1103 _Interface_: 'subroutine omp_destroy_nest_lock(nvar)' 1104 'integer(omp_nest_lock_kind), intent(inout) :: nvar' 1105 1106 _See also_: 1107 *note omp_init_lock:: 1108 1109 _Reference_: 1110 OpenMP specification v4.5 (https://www.openmp.org), Section 3.3.3. 1111 1112 1113 File: libgomp.info, Node: omp_get_wtick, Next: omp_get_wtime, Prev: omp_destroy_nest_lock, Up: Runtime Library Routines 1114 1115 2.40 'omp_get_wtick' - Get timer precision 1116 ========================================== 1117 1118 _Description_: 1119 Gets the timer precision, i.e., the number of seconds between two 1120 successive clock ticks. 1121 1122 _C/C++_: 1123 _Prototype_: 'double omp_get_wtick(void);' 1124 1125 _Fortran_: 1126 _Interface_: 'double precision function omp_get_wtick()' 1127 1128 _See also_: 1129 *note omp_get_wtime:: 1130 1131 _Reference_: 1132 OpenMP specification v4.5 (https://www.openmp.org), Section 3.4.2. 1133 1134 1135 File: libgomp.info, Node: omp_get_wtime, Prev: omp_get_wtick, Up: Runtime Library Routines 1136 1137 2.41 'omp_get_wtime' - Elapsed wall clock time 1138 ============================================== 1139 1140 _Description_: 1141 Elapsed wall clock time in seconds. The time is measured per 1142 thread, no guarantee can be made that two distinct threads measure 1143 the same time. Time is measured from some "time in the past", 1144 which is an arbitrary time guaranteed not to change during the 1145 execution of the program. 1146 1147 _C/C++_: 1148 _Prototype_: 'double omp_get_wtime(void);' 1149 1150 _Fortran_: 1151 _Interface_: 'double precision function omp_get_wtime()' 1152 1153 _See also_: 1154 *note omp_get_wtick:: 1155 1156 _Reference_: 1157 OpenMP specification v4.5 (https://www.openmp.org), Section 3.4.1. 1158 1159 1160 File: libgomp.info, Node: Environment Variables, Next: Enabling OpenACC, Prev: Runtime Library Routines, Up: Top 1161 1162 3 OpenMP Environment Variables 1163 ****************************** 1164 1165 The environment variables which beginning with 'OMP_' are defined by 1166 section 4 of the OpenMP specification in version 4.5, while those 1167 beginning with 'GOMP_' are GNU extensions. 1168 1169 * Menu: 1170 1171 * OMP_CANCELLATION:: Set whether cancellation is activated 1172 * OMP_DISPLAY_ENV:: Show OpenMP version and environment variables 1173 * OMP_DEFAULT_DEVICE:: Set the device used in target regions 1174 * OMP_DYNAMIC:: Dynamic adjustment of threads 1175 * OMP_MAX_ACTIVE_LEVELS:: Set the maximum number of nested parallel regions 1176 * OMP_MAX_TASK_PRIORITY:: Set the maximum task priority value 1177 * OMP_NESTED:: Nested parallel regions 1178 * OMP_NUM_THREADS:: Specifies the number of threads to use 1179 * OMP_PROC_BIND:: Whether theads may be moved between CPUs 1180 * OMP_PLACES:: Specifies on which CPUs the theads should be placed 1181 * OMP_STACKSIZE:: Set default thread stack size 1182 * OMP_SCHEDULE:: How threads are scheduled 1183 * OMP_THREAD_LIMIT:: Set the maximum number of threads 1184 * OMP_WAIT_POLICY:: How waiting threads are handled 1185 * GOMP_CPU_AFFINITY:: Bind threads to specific CPUs 1186 * GOMP_DEBUG:: Enable debugging output 1187 * GOMP_STACKSIZE:: Set default thread stack size 1188 * GOMP_SPINCOUNT:: Set the busy-wait spin count 1189 * GOMP_RTEMS_THREAD_POOLS:: Set the RTEMS specific thread pools 1190 1191 1192 File: libgomp.info, Node: OMP_CANCELLATION, Next: OMP_DISPLAY_ENV, Up: Environment Variables 1193 1194 3.1 'OMP_CANCELLATION' - Set whether cancellation is activated 1195 ============================================================== 1196 1197 _Description_: 1198 If set to 'TRUE', the cancellation is activated. If set to 'FALSE' 1199 or if unset, cancellation is disabled and the 'cancel' construct is 1200 ignored. 1201 1202 _See also_: 1203 *note omp_get_cancellation:: 1204 1205 _Reference_: 1206 OpenMP specification v4.5 (https://www.openmp.org), Section 4.11 1207 1208 1209 File: libgomp.info, Node: OMP_DISPLAY_ENV, Next: OMP_DEFAULT_DEVICE, Prev: OMP_CANCELLATION, Up: Environment Variables 1210 1211 3.2 'OMP_DISPLAY_ENV' - Show OpenMP version and environment variables 1212 ===================================================================== 1213 1214 _Description_: 1215 If set to 'TRUE', the OpenMP version number and the values 1216 associated with the OpenMP environment variables are printed to 1217 'stderr'. If set to 'VERBOSE', it additionally shows the value of 1218 the environment variables which are GNU extensions. If undefined 1219 or set to 'FALSE', this information will not be shown. 1220 1221 _Reference_: 1222 OpenMP specification v4.5 (https://www.openmp.org), Section 4.12 1223 1224 1225 File: libgomp.info, Node: OMP_DEFAULT_DEVICE, Next: OMP_DYNAMIC, Prev: OMP_DISPLAY_ENV, Up: Environment Variables 1226 1227 3.3 'OMP_DEFAULT_DEVICE' - Set the device used in target regions 1228 ================================================================ 1229 1230 _Description_: 1231 Set to choose the device which is used in a 'target' region, unless 1232 the value is overridden by 'omp_set_default_device' or by a 1233 'device' clause. The value shall be the nonnegative device number. 1234 If no device with the given device number exists, the code is 1235 executed on the host. If unset, device number 0 will be used. 1236 1237 _See also_: 1238 *note omp_get_default_device::, *note omp_set_default_device::, 1239 1240 _Reference_: 1241 OpenMP specification v4.5 (https://www.openmp.org), Section 4.13 1242 1243 1244 File: libgomp.info, Node: OMP_DYNAMIC, Next: OMP_MAX_ACTIVE_LEVELS, Prev: OMP_DEFAULT_DEVICE, Up: Environment Variables 1245 1246 3.4 'OMP_DYNAMIC' - Dynamic adjustment of threads 1247 ================================================= 1248 1249 _Description_: 1250 Enable or disable the dynamic adjustment of the number of threads 1251 within a team. The value of this environment variable shall be 1252 'TRUE' or 'FALSE'. If undefined, dynamic adjustment is disabled by 1253 default. 1254 1255 _See also_: 1256 *note omp_set_dynamic:: 1257 1258 _Reference_: 1259 OpenMP specification v4.5 (https://www.openmp.org), Section 4.3 1260 1261 1262 File: libgomp.info, Node: OMP_MAX_ACTIVE_LEVELS, Next: OMP_MAX_TASK_PRIORITY, Prev: OMP_DYNAMIC, Up: Environment Variables 1263 1264 3.5 'OMP_MAX_ACTIVE_LEVELS' - Set the maximum number of nested parallel regions 1265 =============================================================================== 1266 1267 _Description_: 1268 Specifies the initial value for the maximum number of nested 1269 parallel regions. The value of this variable shall be a positive 1270 integer. If undefined, the number of active levels is unlimited. 1271 1272 _See also_: 1273 *note omp_set_max_active_levels:: 1274 1275 _Reference_: 1276 OpenMP specification v4.5 (https://www.openmp.org), Section 4.9 1277 1278 1279 File: libgomp.info, Node: OMP_MAX_TASK_PRIORITY, Next: OMP_NESTED, Prev: OMP_MAX_ACTIVE_LEVELS, Up: Environment Variables 1280 1281 3.6 'OMP_MAX_TASK_PRIORITY' - Set the maximum priority 1282 ====================================================== 1283 1284 number that can be set for a task. 1285 _Description_: 1286 Specifies the initial value for the maximum priority value that can 1287 be set for a task. The value of this variable shall be a 1288 non-negative integer, and zero is allowed. If undefined, the 1289 default priority is 0. 1290 1291 _See also_: 1292 *note omp_get_max_task_priority:: 1293 1294 _Reference_: 1295 OpenMP specification v4.5 (https://www.openmp.org), Section 4.14 1296 1297 1298 File: libgomp.info, Node: OMP_NESTED, Next: OMP_NUM_THREADS, Prev: OMP_MAX_TASK_PRIORITY, Up: Environment Variables 1299 1300 3.7 'OMP_NESTED' - Nested parallel regions 1301 ========================================== 1302 1303 _Description_: 1304 Enable or disable nested parallel regions, i.e., whether team 1305 members are allowed to create new teams. The value of this 1306 environment variable shall be 'TRUE' or 'FALSE'. If undefined, 1307 nested parallel regions are disabled by default. 1308 1309 _See also_: 1310 *note omp_set_nested:: 1311 1312 _Reference_: 1313 OpenMP specification v4.5 (https://www.openmp.org), Section 4.6 1314 1315 1316 File: libgomp.info, Node: OMP_NUM_THREADS, Next: OMP_PROC_BIND, Prev: OMP_NESTED, Up: Environment Variables 1317 1318 3.8 'OMP_NUM_THREADS' - Specifies the number of threads to use 1319 ============================================================== 1320 1321 _Description_: 1322 Specifies the default number of threads to use in parallel regions. 1323 The value of this variable shall be a comma-separated list of 1324 positive integers; the value specified the number of threads to use 1325 for the corresponding nested level. If undefined one thread per 1326 CPU is used. 1327 1328 _See also_: 1329 *note omp_set_num_threads:: 1330 1331 _Reference_: 1332 OpenMP specification v4.5 (https://www.openmp.org), Section 4.2 1333 1334 1335 File: libgomp.info, Node: OMP_PROC_BIND, Next: OMP_PLACES, Prev: OMP_NUM_THREADS, Up: Environment Variables 1336 1337 3.9 'OMP_PROC_BIND' - Whether theads may be moved between CPUs 1338 ============================================================== 1339 1340 _Description_: 1341 Specifies whether threads may be moved between processors. If set 1342 to 'TRUE', OpenMP theads should not be moved; if set to 'FALSE' 1343 they may be moved. Alternatively, a comma separated list with the 1344 values 'MASTER', 'CLOSE' and 'SPREAD' can be used to specify the 1345 thread affinity policy for the corresponding nesting level. With 1346 'MASTER' the worker threads are in the same place partition as the 1347 master thread. With 'CLOSE' those are kept close to the master 1348 thread in contiguous place partitions. And with 'SPREAD' a sparse 1349 distribution across the place partitions is used. 1350 1351 When undefined, 'OMP_PROC_BIND' defaults to 'TRUE' when 1352 'OMP_PLACES' or 'GOMP_CPU_AFFINITY' is set and 'FALSE' otherwise. 1353 1354 _See also_: 1355 *note OMP_PLACES::, *note GOMP_CPU_AFFINITY::, *note 1356 omp_get_proc_bind:: 1357 1358 _Reference_: 1359 OpenMP specification v4.5 (https://www.openmp.org), Section 4.4 1360 1361 1362 File: libgomp.info, Node: OMP_PLACES, Next: OMP_STACKSIZE, Prev: OMP_PROC_BIND, Up: Environment Variables 1363 1364 3.10 'OMP_PLACES' - Specifies on which CPUs the theads should be placed 1365 ======================================================================= 1366 1367 _Description_: 1368 The thread placement can be either specified using an abstract name 1369 or by an explicit list of the places. The abstract names 1370 'threads', 'cores' and 'sockets' can be optionally followed by a 1371 positive number in parentheses, which denotes the how many places 1372 shall be created. With 'threads' each place corresponds to a 1373 single hardware thread; 'cores' to a single core with the 1374 corresponding number of hardware threads; and with 'sockets' the 1375 place corresponds to a single socket. The resulting placement can 1376 be shown by setting the 'OMP_DISPLAY_ENV' environment variable. 1377 1378 Alternatively, the placement can be specified explicitly as 1379 comma-separated list of places. A place is specified by set of 1380 nonnegative numbers in curly braces, denoting the denoting the 1381 hardware threads. The hardware threads belonging to a place can 1382 either be specified as comma-separated list of nonnegative thread 1383 numbers or using an interval. Multiple places can also be either 1384 specified by a comma-separated list of places or by an interval. 1385 To specify an interval, a colon followed by the count is placed 1386 after after the hardware thread number or the place. Optionally, 1387 the length can be followed by a colon and the stride number - 1388 otherwise a unit stride is assumed. For instance, the following 1389 specifies the same places list: '"{0,1,2}, {3,4,6}, {7,8,9}, 1390 {10,11,12}"'; '"{0:3}, {3:3}, {7:3}, {10:3}"'; and '"{0:2}:4:3"'. 1391 1392 If 'OMP_PLACES' and 'GOMP_CPU_AFFINITY' are unset and 1393 'OMP_PROC_BIND' is either unset or 'false', threads may be moved 1394 between CPUs following no placement policy. 1395 1396 _See also_: 1397 *note OMP_PROC_BIND::, *note GOMP_CPU_AFFINITY::, *note 1398 omp_get_proc_bind::, *note OMP_DISPLAY_ENV:: 1399 1400 _Reference_: 1401 OpenMP specification v4.5 (https://www.openmp.org), Section 4.5 1402 1403 1404 File: libgomp.info, Node: OMP_STACKSIZE, Next: OMP_SCHEDULE, Prev: OMP_PLACES, Up: Environment Variables 1405 1406 3.11 'OMP_STACKSIZE' - Set default thread stack size 1407 ==================================================== 1408 1409 _Description_: 1410 Set the default thread stack size in kilobytes, unless the number 1411 is suffixed by 'B', 'K', 'M' or 'G', in which case the size is, 1412 respectively, in bytes, kilobytes, megabytes or gigabytes. This is 1413 different from 'pthread_attr_setstacksize' which gets the number of 1414 bytes as an argument. If the stack size cannot be set due to 1415 system constraints, an error is reported and the initial stack size 1416 is left unchanged. If undefined, the stack size is system 1417 dependent. 1418 1419 _Reference_: 1420 OpenMP specification v4.5 (https://www.openmp.org), Section 4.7 1421 1422 1423 File: libgomp.info, Node: OMP_SCHEDULE, Next: OMP_THREAD_LIMIT, Prev: OMP_STACKSIZE, Up: Environment Variables 1424 1425 3.12 'OMP_SCHEDULE' - How threads are scheduled 1426 =============================================== 1427 1428 _Description_: 1429 Allows to specify 'schedule type' and 'chunk size'. The value of 1430 the variable shall have the form: 'type[,chunk]' where 'type' is 1431 one of 'static', 'dynamic', 'guided' or 'auto' The optional 'chunk' 1432 size shall be a positive integer. If undefined, dynamic scheduling 1433 and a chunk size of 1 is used. 1434 1435 _See also_: 1436 *note omp_set_schedule:: 1437 1438 _Reference_: 1439 OpenMP specification v4.5 (https://www.openmp.org), Sections 1440 2.7.1.1 and 4.1 1441 1442 1443 File: libgomp.info, Node: OMP_THREAD_LIMIT, Next: OMP_WAIT_POLICY, Prev: OMP_SCHEDULE, Up: Environment Variables 1444 1445 3.13 'OMP_THREAD_LIMIT' - Set the maximum number of threads 1446 =========================================================== 1447 1448 _Description_: 1449 Specifies the number of threads to use for the whole program. The 1450 value of this variable shall be a positive integer. If undefined, 1451 the number of threads is not limited. 1452 1453 _See also_: 1454 *note OMP_NUM_THREADS::, *note omp_get_thread_limit:: 1455 1456 _Reference_: 1457 OpenMP specification v4.5 (https://www.openmp.org), Section 4.10 1458 1459 1460 File: libgomp.info, Node: OMP_WAIT_POLICY, Next: GOMP_CPU_AFFINITY, Prev: OMP_THREAD_LIMIT, Up: Environment Variables 1461 1462 3.14 'OMP_WAIT_POLICY' - How waiting threads are handled 1463 ======================================================== 1464 1465 _Description_: 1466 Specifies whether waiting threads should be active or passive. If 1467 the value is 'PASSIVE', waiting threads should not consume CPU 1468 power while waiting; while the value is 'ACTIVE' specifies that 1469 they should. If undefined, threads wait actively for a short time 1470 before waiting passively. 1471 1472 _See also_: 1473 *note GOMP_SPINCOUNT:: 1474 1475 _Reference_: 1476 OpenMP specification v4.5 (https://www.openmp.org), Section 4.8 1477 1478 1479 File: libgomp.info, Node: GOMP_CPU_AFFINITY, Next: GOMP_DEBUG, Prev: OMP_WAIT_POLICY, Up: Environment Variables 1480 1481 3.15 'GOMP_CPU_AFFINITY' - Bind threads to specific CPUs 1482 ======================================================== 1483 1484 _Description_: 1485 Binds threads to specific CPUs. The variable should contain a 1486 space-separated or comma-separated list of CPUs. This list may 1487 contain different kinds of entries: either single CPU numbers in 1488 any order, a range of CPUs (M-N) or a range with some stride 1489 (M-N:S). CPU numbers are zero based. For example, 1490 'GOMP_CPU_AFFINITY="0 3 1-2 4-15:2"' will bind the initial thread 1491 to CPU 0, the second to CPU 3, the third to CPU 1, the fourth to 1492 CPU 2, the fifth to CPU 4, the sixth through tenth to CPUs 6, 8, 1493 10, 12, and 14 respectively and then start assigning back from the 1494 beginning of the list. 'GOMP_CPU_AFFINITY=0' binds all threads to 1495 CPU 0. 1496 1497 There is no libgomp library routine to determine whether a CPU 1498 affinity specification is in effect. As a workaround, 1499 language-specific library functions, e.g., 'getenv' in C or 1500 'GET_ENVIRONMENT_VARIABLE' in Fortran, may be used to query the 1501 setting of the 'GOMP_CPU_AFFINITY' environment variable. A defined 1502 CPU affinity on startup cannot be changed or disabled during the 1503 runtime of the application. 1504 1505 If both 'GOMP_CPU_AFFINITY' and 'OMP_PROC_BIND' are set, 1506 'OMP_PROC_BIND' has a higher precedence. If neither has been set 1507 and 'OMP_PROC_BIND' is unset, or when 'OMP_PROC_BIND' is set to 1508 'FALSE', the host system will handle the assignment of threads to 1509 CPUs. 1510 1511 _See also_: 1512 *note OMP_PLACES::, *note OMP_PROC_BIND:: 1513 1514 1515 File: libgomp.info, Node: GOMP_DEBUG, Next: GOMP_STACKSIZE, Prev: GOMP_CPU_AFFINITY, Up: Environment Variables 1516 1517 3.16 'GOMP_DEBUG' - Enable debugging output 1518 =========================================== 1519 1520 _Description_: 1521 Enable debugging output. The variable should be set to '0' 1522 (disabled, also the default if not set), or '1' (enabled). 1523 1524 If enabled, some debugging output will be printed during execution. 1525 This is currently not specified in more detail, and subject to 1526 change. 1527 1528 1529 File: libgomp.info, Node: GOMP_STACKSIZE, Next: GOMP_SPINCOUNT, Prev: GOMP_DEBUG, Up: Environment Variables 1530 1531 3.17 'GOMP_STACKSIZE' - Set default thread stack size 1532 ===================================================== 1533 1534 _Description_: 1535 Set the default thread stack size in kilobytes. This is different 1536 from 'pthread_attr_setstacksize' which gets the number of bytes as 1537 an argument. If the stack size cannot be set due to system 1538 constraints, an error is reported and the initial stack size is 1539 left unchanged. If undefined, the stack size is system dependent. 1540 1541 _See also_: 1542 *note OMP_STACKSIZE:: 1543 1544 _Reference_: 1545 GCC Patches Mailinglist 1546 (http://gcc.gnu.org/ml/gcc-patches/2006-06/msg00493.html), GCC 1547 Patches Mailinglist 1548 (http://gcc.gnu.org/ml/gcc-patches/2006-06/msg00496.html) 1549 1550 1551 File: libgomp.info, Node: GOMP_SPINCOUNT, Next: GOMP_RTEMS_THREAD_POOLS, Prev: GOMP_STACKSIZE, Up: Environment Variables 1552 1553 3.18 'GOMP_SPINCOUNT' - Set the busy-wait spin count 1554 ==================================================== 1555 1556 _Description_: 1557 Determines how long a threads waits actively with consuming CPU 1558 power before waiting passively without consuming CPU power. The 1559 value may be either 'INFINITE', 'INFINITY' to always wait actively 1560 or an integer which gives the number of spins of the busy-wait 1561 loop. The integer may optionally be followed by the following 1562 suffixes acting as multiplication factors: 'k' (kilo, thousand), 1563 'M' (mega, million), 'G' (giga, billion), or 'T' (tera, trillion). 1564 If undefined, 0 is used when 'OMP_WAIT_POLICY' is 'PASSIVE', 1565 300,000 is used when 'OMP_WAIT_POLICY' is undefined and 30 billion 1566 is used when 'OMP_WAIT_POLICY' is 'ACTIVE'. If there are more 1567 OpenMP threads than available CPUs, 1000 and 100 spins are used for 1568 'OMP_WAIT_POLICY' being 'ACTIVE' or undefined, respectively; unless 1569 the 'GOMP_SPINCOUNT' is lower or 'OMP_WAIT_POLICY' is 'PASSIVE'. 1570 1571 _See also_: 1572 *note OMP_WAIT_POLICY:: 1573 1574 1575 File: libgomp.info, Node: GOMP_RTEMS_THREAD_POOLS, Prev: GOMP_SPINCOUNT, Up: Environment Variables 1576 1577 3.19 'GOMP_RTEMS_THREAD_POOLS' - Set the RTEMS specific thread pools 1578 ==================================================================== 1579 1580 _Description_: 1581 This environment variable is only used on the RTEMS real-time 1582 operating system. It determines the scheduler instance specific 1583 thread pools. The format for 'GOMP_RTEMS_THREAD_POOLS' is a list 1584 of optional '<thread-pool-count>[$<priority>]@<scheduler-name>' 1585 configurations separated by ':' where: 1586 * '<thread-pool-count>' is the thread pool count for this 1587 scheduler instance. 1588 * '$<priority>' is an optional priority for the worker threads 1589 of a thread pool according to 'pthread_setschedparam'. In 1590 case a priority value is omitted, then a worker thread will 1591 inherit the priority of the OpenMP master thread that created 1592 it. The priority of the worker thread is not changed after 1593 creation, even if a new OpenMP master thread using the worker 1594 has a different priority. 1595 * '@<scheduler-name>' is the scheduler instance name according 1596 to the RTEMS application configuration. 1597 In case no thread pool configuration is specified for a scheduler 1598 instance, then each OpenMP master thread of this scheduler instance 1599 will use its own dynamically allocated thread pool. To limit the 1600 worker thread count of the thread pools, each OpenMP master thread 1601 must call 'omp_set_num_threads'. 1602 _Example_: 1603 Lets suppose we have three scheduler instances 'IO', 'WRK0', and 1604 'WRK1' with 'GOMP_RTEMS_THREAD_POOLS' set to '"1@WRK0:3$4@WRK1"'. 1605 Then there are no thread pool restrictions for scheduler instance 1606 'IO'. In the scheduler instance 'WRK0' there is one thread pool 1607 available. Since no priority is specified for this scheduler 1608 instance, the worker thread inherits the priority of the OpenMP 1609 master thread that created it. In the scheduler instance 'WRK1' 1610 there are three thread pools available and their worker threads run 1611 at priority four. 1612 1613 1614 File: libgomp.info, Node: Enabling OpenACC, Next: OpenACC Runtime Library Routines, Prev: Environment Variables, Up: Top 1615 1616 4 Enabling OpenACC 1617 ****************** 1618 1619 To activate the OpenACC extensions for C/C++ and Fortran, the 1620 compile-time flag '-fopenacc' must be specified. This enables the 1621 OpenACC directive '#pragma acc' in C/C++ and '!$accp' directives in free 1622 form, 'c$acc', '*$acc' and '!$acc' directives in fixed form, '!$' 1623 conditional compilation sentinels in free form and 'c$', '*$' and '!$' 1624 sentinels in fixed form, for Fortran. The flag also arranges for 1625 automatic linking of the OpenACC runtime library (*note OpenACC Runtime 1626 Library Routines::). 1627 1628 A complete description of all OpenACC directives accepted may be 1629 found in the OpenACC (https://www.openacc.org) Application Programming 1630 Interface manual, version 2.0. 1631 1632 Note that this is an experimental feature and subject to change in 1633 future versions of GCC. See <https://gcc.gnu.org/wiki/OpenACC> for more 1634 information. 1635 1636 1637 File: libgomp.info, Node: OpenACC Runtime Library Routines, Next: OpenACC Environment Variables, Prev: Enabling OpenACC, Up: Top 1638 1639 5 OpenACC Runtime Library Routines 1640 ********************************** 1641 1642 The runtime routines described here are defined by section 3 of the 1643 OpenACC specifications in version 2.0. They have C linkage, and do not 1644 throw exceptions. Generally, they are available only for the host, with 1645 the exception of 'acc_on_device', which is available for both the host 1646 and the acceleration device. 1647 1648 * Menu: 1649 1650 * acc_get_num_devices:: Get number of devices for the given device 1651 type. 1652 * acc_set_device_type:: Set type of device accelerator to use. 1653 * acc_get_device_type:: Get type of device accelerator to be used. 1654 * acc_set_device_num:: Set device number to use. 1655 * acc_get_device_num:: Get device number to be used. 1656 * acc_async_test:: Tests for completion of a specific asynchronous 1657 operation. 1658 * acc_async_test_all:: Tests for completion of all asychronous 1659 operations. 1660 * acc_wait:: Wait for completion of a specific asynchronous 1661 operation. 1662 * acc_wait_all:: Waits for completion of all asyncrhonous 1663 operations. 1664 * acc_wait_all_async:: Wait for completion of all asynchronous 1665 operations. 1666 * acc_wait_async:: Wait for completion of asynchronous operations. 1667 * acc_init:: Initialize runtime for a specific device type. 1668 * acc_shutdown:: Shuts down the runtime for a specific device 1669 type. 1670 * acc_on_device:: Whether executing on a particular device 1671 * acc_malloc:: Allocate device memory. 1672 * acc_free:: Free device memory. 1673 * acc_copyin:: Allocate device memory and copy host memory to 1674 it. 1675 * acc_present_or_copyin:: If the data is not present on the device, 1676 allocate device memory and copy from host 1677 memory. 1678 * acc_create:: Allocate device memory and map it to host 1679 memory. 1680 * acc_present_or_create:: If the data is not present on the device, 1681 allocate device memory and map it to host 1682 memory. 1683 * acc_copyout:: Copy device memory to host memory. 1684 * acc_delete:: Free device memory. 1685 * acc_update_device:: Update device memory from mapped host memory. 1686 * acc_update_self:: Update host memory from mapped device memory. 1687 * acc_map_data:: Map previously allocated device memory to host 1688 memory. 1689 * acc_unmap_data:: Unmap device memory from host memory. 1690 * acc_deviceptr:: Get device pointer associated with specific 1691 host address. 1692 * acc_hostptr:: Get host pointer associated with specific 1693 device address. 1694 * acc_is_present:: Indiciate whether host variable / array is 1695 present on device. 1696 * acc_memcpy_to_device:: Copy host memory to device memory. 1697 * acc_memcpy_from_device:: Copy device memory to host memory. 1698 1699 API routines for target platforms. 1700 1701 * acc_get_current_cuda_device:: Get CUDA device handle. 1702 * acc_get_current_cuda_context::Get CUDA context handle. 1703 * acc_get_cuda_stream:: Get CUDA stream handle. 1704 * acc_set_cuda_stream:: Set CUDA stream handle. 1705 1706 1707 File: libgomp.info, Node: acc_get_num_devices, Next: acc_set_device_type, Up: OpenACC Runtime Library Routines 1708 1709 5.1 'acc_get_num_devices' - Get number of devices for given device type 1710 ======================================================================= 1711 1712 _Description_ 1713 This function returns a value indicating the number of devices 1714 available for the device type specified in DEVICETYPE. 1715 1716 _C/C++_: 1717 _Prototype_: 'int acc_get_num_devices(acc_device_t devicetype);' 1718 1719 _Fortran_: 1720 _Interface_: 'integer function acc_get_num_devices(devicetype)' 1721 'integer(kind=acc_device_kind) devicetype' 1722 1723 _Reference_: 1724 OpenACC specification v2.0 (https://www.openacc.org), section 1725 3.2.1. 1726 1727 1728 File: libgomp.info, Node: acc_set_device_type, Next: acc_get_device_type, Prev: acc_get_num_devices, Up: OpenACC Runtime Library Routines 1729 1730 5.2 'acc_set_device_type' - Set type of device accelerator to use. 1731 ================================================================== 1732 1733 _Description_ 1734 This function indicates to the runtime library which device typr, 1735 specified in DEVICETYPE, to use when executing a parallel or 1736 kernels region. 1737 1738 _C/C++_: 1739 _Prototype_: 'acc_set_device_type(acc_device_t devicetype);' 1740 1741 _Fortran_: 1742 _Interface_: 'subroutine acc_set_device_type(devicetype)' 1743 'integer(kind=acc_device_kind) devicetype' 1744 1745 _Reference_: 1746 OpenACC specification v2.0 (https://www.openacc.org), section 1747 3.2.2. 1748 1749 1750 File: libgomp.info, Node: acc_get_device_type, Next: acc_set_device_num, Prev: acc_set_device_type, Up: OpenACC Runtime Library Routines 1751 1752 5.3 'acc_get_device_type' - Get type of device accelerator to be used. 1753 ====================================================================== 1754 1755 _Description_ 1756 This function returns what device type will be used when executing 1757 a parallel or kernels region. 1758 1759 _C/C++_: 1760 _Prototype_: 'acc_device_t acc_get_device_type(void);' 1761 1762 _Fortran_: 1763 _Interface_: 'function acc_get_device_type(void)' 1764 'integer(kind=acc_device_kind) acc_get_device_type' 1765 1766 _Reference_: 1767 OpenACC specification v2.0 (https://www.openacc.org), section 1768 3.2.3. 1769 1770 1771 File: libgomp.info, Node: acc_set_device_num, Next: acc_get_device_num, Prev: acc_get_device_type, Up: OpenACC Runtime Library Routines 1772 1773 5.4 'acc_set_device_num' - Set device number to use. 1774 ==================================================== 1775 1776 _Description_ 1777 This function will indicate to the runtime which device number, 1778 specified by NUM, associated with the specifed device type 1779 DEVICETYPE. 1780 1781 _C/C++_: 1782 _Prototype_: 'acc_set_device_num(int num, acc_device_t devicetype);' 1783 1784 _Fortran_: 1785 _Interface_: 'subroutine acc_set_device_num(devicenum, devicetype)' 1786 'integer devicenum' 1787 'integer(kind=acc_device_kind) devicetype' 1788 1789 _Reference_: 1790 OpenACC specification v2.0 (https://www.openacc.org), section 1791 3.2.4. 1792 1793 1794 File: libgomp.info, Node: acc_get_device_num, Next: acc_async_test, Prev: acc_set_device_num, Up: OpenACC Runtime Library Routines 1795 1796 5.5 'acc_get_device_num' - Get device number to be used. 1797 ======================================================== 1798 1799 _Description_ 1800 This function returns which device number associated with the 1801 specified device type DEVICETYPE, will be used when executing a 1802 parallel or kernels region. 1803 1804 _C/C++_: 1805 _Prototype_: 'int acc_get_device_num(acc_device_t devicetype);' 1806 1807 _Fortran_: 1808 _Interface_: 'function acc_get_device_num(devicetype)' 1809 'integer(kind=acc_device_kind) devicetype' 1810 'integer acc_get_device_num' 1811 1812 _Reference_: 1813 OpenACC specification v2.0 (https://www.openacc.org), section 1814 3.2.5. 1815 1816 1817 File: libgomp.info, Node: acc_async_test, Next: acc_async_test_all, Prev: acc_get_device_num, Up: OpenACC Runtime Library Routines 1818 1819 5.6 'acc_async_test' - Test for completion of a specific asynchronous operation. 1820 ================================================================================ 1821 1822 _Description_ 1823 This function tests for completion of the asynchrounous operation 1824 specified in ARG. In C/C++, a non-zero value will be returned to 1825 indicate the specified asynchronous operation has completed. While 1826 Fortran will return a 'true'. If the asynchrounous operation has 1827 not completed, C/C++ returns a zero and Fortran returns a 'false'. 1828 1829 _C/C++_: 1830 _Prototype_: 'int acc_async_test(int arg);' 1831 1832 _Fortran_: 1833 _Interface_: 'function acc_async_test(arg)' 1834 'integer(kind=acc_handle_kind) arg' 1835 'logical acc_async_test' 1836 1837 _Reference_: 1838 OpenACC specification v2.0 (https://www.openacc.org), section 1839 3.2.6. 1840 1841 1842 File: libgomp.info, Node: acc_async_test_all, Next: acc_wait, Prev: acc_async_test, Up: OpenACC Runtime Library Routines 1843 1844 5.7 'acc_async_test_all' - Tests for completion of all asynchronous operations. 1845 =============================================================================== 1846 1847 _Description_ 1848 This function tests for completion of all asynchrounous operations. 1849 In C/C++, a non-zero value will be returned to indicate all 1850 asynchronous operations have completed. While Fortran will return 1851 a 'true'. If any asynchronous operation has not completed, C/C++ 1852 returns a zero and Fortran returns a 'false'. 1853 1854 _C/C++_: 1855 _Prototype_: 'int acc_async_test_all(void);' 1856 1857 _Fortran_: 1858 _Interface_: 'function acc_async_test()' 1859 'logical acc_get_device_num' 1860 1861 _Reference_: 1862 OpenACC specification v2.0 (https://www.openacc.org), section 1863 3.2.7. 1864 1865 1866 File: libgomp.info, Node: acc_wait, Next: acc_wait_all, Prev: acc_async_test_all, Up: OpenACC Runtime Library Routines 1867 1868 5.8 'acc_wait' - Wait for completion of a specific asynchronous operation. 1869 ========================================================================== 1870 1871 _Description_ 1872 This function waits for completion of the asynchronous operation 1873 specified in ARG. 1874 1875 _C/C++_: 1876 _Prototype_: 'acc_wait(arg);' 1877 _Prototype 'acc_async_wait(arg);' 1878 (OpenACC 1.0 1879 compatibility)_: 1880 1881 _Fortran_: 1882 _Interface_: 'subroutine acc_wait(arg)' 1883 'integer(acc_handle_kind) arg' 1884 _Interface 'subroutine acc_async_wait(arg)' 1885 (OpenACC 1.0 1886 compatibility)_: 1887 'integer(acc_handle_kind) arg' 1888 1889 _Reference_: 1890 OpenACC specification v2.0 (https://www.openacc.org), section 1891 3.2.8. 1892 1893 1894 File: libgomp.info, Node: acc_wait_all, Next: acc_wait_all_async, Prev: acc_wait, Up: OpenACC Runtime Library Routines 1895 1896 5.9 'acc_wait_all' - Waits for completion of all asynchronous operations. 1897 ========================================================================= 1898 1899 _Description_ 1900 This function waits for the completion of all asynchronous 1901 operations. 1902 1903 _C/C++_: 1904 _Prototype_: 'acc_wait_all(void);' 1905 _Prototype 'acc_async_wait_all(void);' 1906 (OpenACC 1.0 1907 compatibility)_: 1908 1909 _Fortran_: 1910 _Interface_: 'subroutine acc_wait_all()' 1911 _Interface 'subroutine acc_async_wait_all()' 1912 (OpenACC 1.0 1913 compatibility)_: 1914 1915 _Reference_: 1916 OpenACC specification v2.0 (https://www.openacc.org), section 1917 3.2.10. 1918 1919 1920 File: libgomp.info, Node: acc_wait_all_async, Next: acc_wait_async, Prev: acc_wait_all, Up: OpenACC Runtime Library Routines 1921 1922 5.10 'acc_wait_all_async' - Wait for completion of all asynchronous operations. 1923 =============================================================================== 1924 1925 _Description_ 1926 This function enqueues a wait operation on the queue ASYNC for any 1927 and all asynchronous operations that have been previously enqueued 1928 on any queue. 1929 1930 _C/C++_: 1931 _Prototype_: 'acc_wait_all_async(int async);' 1932 1933 _Fortran_: 1934 _Interface_: 'subroutine acc_wait_all_async(async)' 1935 'integer(acc_handle_kind) async' 1936 1937 _Reference_: 1938 OpenACC specification v2.0 (https://www.openacc.org), section 1939 3.2.11. 1940 1941 1942 File: libgomp.info, Node: acc_wait_async, Next: acc_init, Prev: acc_wait_all_async, Up: OpenACC Runtime Library Routines 1943 1944 5.11 'acc_wait_async' - Wait for completion of asynchronous operations. 1945 ======================================================================= 1946 1947 _Description_ 1948 This function enqueues a wait operation on queue ASYNC for any and 1949 all asynchronous operations enqueued on queue ARG. 1950 1951 _C/C++_: 1952 _Prototype_: 'acc_wait_async(int arg, int async);' 1953 1954 _Fortran_: 1955 _Interface_: 'subroutine acc_wait_async(arg, async)' 1956 'integer(acc_handle_kind) arg, async' 1957 1958 _Reference_: 1959 OpenACC specification v2.0 (https://www.openacc.org), section 1960 3.2.9. 1961 1962 1963 File: libgomp.info, Node: acc_init, Next: acc_shutdown, Prev: acc_wait_async, Up: OpenACC Runtime Library Routines 1964 1965 5.12 'acc_init' - Initialize runtime for a specific device type. 1966 ================================================================ 1967 1968 _Description_ 1969 This function initializes the runtime for the device type specified 1970 in DEVICETYPE. 1971 1972 _C/C++_: 1973 _Prototype_: 'acc_init(acc_device_t devicetype);' 1974 1975 _Fortran_: 1976 _Interface_: 'subroutine acc_init(devicetype)' 1977 'integer(acc_device_kind) devicetype' 1978 1979 _Reference_: 1980 OpenACC specification v2.0 (https://www.openacc.org), section 1981 3.2.12. 1982 1983 1984 File: libgomp.info, Node: acc_shutdown, Next: acc_on_device, Prev: acc_init, Up: OpenACC Runtime Library Routines 1985 1986 5.13 'acc_shutdown' - Shuts down the runtime for a specific device type. 1987 ======================================================================== 1988 1989 _Description_ 1990 This function shuts down the runtime for the device type specified 1991 in DEVICETYPE. 1992 1993 _C/C++_: 1994 _Prototype_: 'acc_shutdown(acc_device_t devicetype);' 1995 1996 _Fortran_: 1997 _Interface_: 'subroutine acc_shutdown(devicetype)' 1998 'integer(acc_device_kind) devicetype' 1999 2000 _Reference_: 2001 OpenACC specification v2.0 (https://www.openacc.org), section 2002 3.2.13. 2003 2004 2005 File: libgomp.info, Node: acc_on_device, Next: acc_malloc, Prev: acc_shutdown, Up: OpenACC Runtime Library Routines 2006 2007 5.14 'acc_on_device' - Whether executing on a particular device 2008 =============================================================== 2009 2010 _Description_: 2011 This function returns whether the program is executing on a 2012 particular device specified in DEVICETYPE. In C/C++ a non-zero 2013 value is returned to indicate the device is execiting on the 2014 specified device type. In Fortran, 'true' will be returned. If 2015 the program is not executing on the specified device type C/C++ 2016 will return a zero, while Fortran will return 'false'. 2017 2018 _C/C++_: 2019 _Prototype_: 'acc_on_device(acc_device_t devicetype);' 2020 2021 _Fortran_: 2022 _Interface_: 'function acc_on_device(devicetype)' 2023 'integer(acc_device_kind) devicetype' 2024 'logical acc_on_device' 2025 2026 _Reference_: 2027 OpenACC specification v2.0 (https://www.openacc.org), section 2028 3.2.14. 2029 2030 2031 File: libgomp.info, Node: acc_malloc, Next: acc_free, Prev: acc_on_device, Up: OpenACC Runtime Library Routines 2032 2033 5.15 'acc_malloc' - Allocate device memory. 2034 =========================================== 2035 2036 _Description_ 2037 This function allocates LEN bytes of device memory. It returns the 2038 device address of the allocated memory. 2039 2040 _C/C++_: 2041 _Prototype_: 'd_void* acc_malloc(size_t len);' 2042 2043 _Reference_: 2044 OpenACC specification v2.0 (https://www.openacc.org), section 2045 3.2.15. 2046 2047 2048 File: libgomp.info, Node: acc_free, Next: acc_copyin, Prev: acc_malloc, Up: OpenACC Runtime Library Routines 2049 2050 5.16 'acc_free' - Free device memory. 2051 ===================================== 2052 2053 _Description_ 2054 Free previously allocated device memory at the device address 'a'. 2055 2056 _C/C++_: 2057 _Prototype_: 'acc_free(d_void *a);' 2058 2059 _Reference_: 2060 OpenACC specification v2.0 (https://www.openacc.org), section 2061 3.2.16. 2062 2063 2064 File: libgomp.info, Node: acc_copyin, Next: acc_present_or_copyin, Prev: acc_free, Up: OpenACC Runtime Library Routines 2065 2066 5.17 'acc_copyin' - Allocate device memory and copy host memory to it. 2067 ====================================================================== 2068 2069 _Description_ 2070 In C/C++, this function allocates LEN bytes of device memory and 2071 maps it to the specified host address in A. The device address of 2072 the newly allocated device memory is returned. 2073 2074 In Fortran, two (2) forms are supported. In the first form, A 2075 specifies a contiguous array section. The second form A specifies 2076 a variable or array element and LEN specifies the length in bytes. 2077 2078 _C/C++_: 2079 _Prototype_: 'void *acc_copyin(h_void *a, size_t len);' 2080 2081 _Fortran_: 2082 _Interface_: 'subroutine acc_copyin(a)' 2083 'type, dimension(:[,:]...) :: a' 2084 _Interface_: 'subroutine acc_copyin(a, len)' 2085 'type, dimension(:[,:]...) :: a' 2086 'integer len' 2087 2088 _Reference_: 2089 OpenACC specification v2.0 (https://www.openacc.org), section 2090 3.2.17. 2091 2092 2093 File: libgomp.info, Node: acc_present_or_copyin, Next: acc_create, Prev: acc_copyin, Up: OpenACC Runtime Library Routines 2094 2095 5.18 'acc_present_or_copyin' - If the data is not present on the device, allocate device memory and copy from host memory. 2096 ========================================================================================================================== 2097 2098 _Description_ 2099 This function tests if the host data specifed by A and of length 2100 LEN is present or not. If it is not present, then device memory 2101 will be allocated and the host memory copied. The device address 2102 of the newly allocated device memory is returned. 2103 2104 In Fortran, two (2) forms are supported. In the first form, A 2105 specifies a contiguous array section. The second form A specifies 2106 a variable or array element and LEN specifies the length in bytes. 2107 2108 _C/C++_: 2109 _Prototype_: 'void *acc_present_or_copyin(h_void *a, size_t len);' 2110 _Prototype_: 'void *acc_pcopyin(h_void *a, size_t len);' 2111 2112 _Fortran_: 2113 _Interface_: 'subroutine acc_present_or_copyin(a)' 2114 'type, dimension(:[,:]...) :: a' 2115 _Interface_: 'subroutine acc_present_or_copyin(a, len)' 2116 'type, dimension(:[,:]...) :: a' 2117 'integer len' 2118 _Interface_: 'subroutine acc_pcopyin(a)' 2119 'type, dimension(:[,:]...) :: a' 2120 _Interface_: 'subroutine acc_pcopyin(a, len)' 2121 'type, dimension(:[,:]...) :: a' 2122 'integer len' 2123 2124 _Reference_: 2125 OpenACC specification v2.0 (https://www.openacc.org), section 2126 3.2.18. 2127 2128 2129 File: libgomp.info, Node: acc_create, Next: acc_present_or_create, Prev: acc_present_or_copyin, Up: OpenACC Runtime Library Routines 2130 2131 5.19 'acc_create' - Allocate device memory and map it to host memory. 2132 ===================================================================== 2133 2134 _Description_ 2135 This function allocates device memory and maps it to host memory 2136 specified by the host address A with a length of LEN bytes. In 2137 C/C++, the function returns the device address of the allocated 2138 device memory. 2139 2140 In Fortran, two (2) forms are supported. In the first form, A 2141 specifies a contiguous array section. The second form A specifies 2142 a variable or array element and LEN specifies the length in bytes. 2143 2144 _C/C++_: 2145 _Prototype_: 'void *acc_create(h_void *a, size_t len);' 2146 2147 _Fortran_: 2148 _Interface_: 'subroutine acc_create(a)' 2149 'type, dimension(:[,:]...) :: a' 2150 _Interface_: 'subroutine acc_create(a, len)' 2151 'type, dimension(:[,:]...) :: a' 2152 'integer len' 2153 2154 _Reference_: 2155 OpenACC specification v2.0 (https://www.openacc.org), section 2156 3.2.19. 2157 2158 2159 File: libgomp.info, Node: acc_present_or_create, Next: acc_copyout, Prev: acc_create, Up: OpenACC Runtime Library Routines 2160 2161 5.20 'acc_present_or_create' - If the data is not present on the device, allocate device memory and map it to host memory. 2162 ========================================================================================================================== 2163 2164 _Description_ 2165 This function tests if the host data specifed by A and of length 2166 LEN is present or not. If it is not present, then device memory 2167 will be allocated and mapped to host memory. In C/C++, the device 2168 address of the newly allocated device memory is returned. 2169 2170 In Fortran, two (2) forms are supported. In the first form, A 2171 specifies a contiguous array section. The second form A specifies 2172 a variable or array element and LEN specifies the length in bytes. 2173 2174 _C/C++_: 2175 _Prototype_: 'void *acc_present_or_create(h_void *a, size_t len)' 2176 _Prototype_: 'void *acc_pcreate(h_void *a, size_t len)' 2177 2178 _Fortran_: 2179 _Interface_: 'subroutine acc_present_or_create(a)' 2180 'type, dimension(:[,:]...) :: a' 2181 _Interface_: 'subroutine acc_present_or_create(a, len)' 2182 'type, dimension(:[,:]...) :: a' 2183 'integer len' 2184 _Interface_: 'subroutine acc_pcreate(a)' 2185 'type, dimension(:[,:]...) :: a' 2186 _Interface_: 'subroutine acc_pcreate(a, len)' 2187 'type, dimension(:[,:]...) :: a' 2188 'integer len' 2189 2190 _Reference_: 2191 OpenACC specification v2.0 (https://www.openacc.org), section 2192 3.2.20. 2193 2194 2195 File: libgomp.info, Node: acc_copyout, Next: acc_delete, Prev: acc_present_or_create, Up: OpenACC Runtime Library Routines 2196 2197 5.21 'acc_copyout' - Copy device memory to host memory. 2198 ======================================================= 2199 2200 _Description_ 2201 This function copies mapped device memory to host memory which is 2202 specified by host address A for a length LEN bytes in C/C++. 2203 2204 In Fortran, two (2) forms are supported. In the first form, A 2205 specifies a contiguous array section. The second form A specifies 2206 a variable or array element and LEN specifies the length in bytes. 2207 2208 _C/C++_: 2209 _Prototype_: 'acc_copyout(h_void *a, size_t len);' 2210 2211 _Fortran_: 2212 _Interface_: 'subroutine acc_copyout(a)' 2213 'type, dimension(:[,:]...) :: a' 2214 _Interface_: 'subroutine acc_copyout(a, len)' 2215 'type, dimension(:[,:]...) :: a' 2216 'integer len' 2217 2218 _Reference_: 2219 OpenACC specification v2.0 (https://www.openacc.org), section 2220 3.2.21. 2221 2222 2223 File: libgomp.info, Node: acc_delete, Next: acc_update_device, Prev: acc_copyout, Up: OpenACC Runtime Library Routines 2224 2225 5.22 'acc_delete' - Free device memory. 2226 ======================================= 2227 2228 _Description_ 2229 This function frees previously allocated device memory specified by 2230 the device address A and the length of LEN bytes. 2231 2232 In Fortran, two (2) forms are supported. In the first form, A 2233 specifies a contiguous array section. The second form A specifies 2234 a variable or array element and LEN specifies the length in bytes. 2235 2236 _C/C++_: 2237 _Prototype_: 'acc_delete(h_void *a, size_t len);' 2238 2239 _Fortran_: 2240 _Interface_: 'subroutine acc_delete(a)' 2241 'type, dimension(:[,:]...) :: a' 2242 _Interface_: 'subroutine acc_delete(a, len)' 2243 'type, dimension(:[,:]...) :: a' 2244 'integer len' 2245 2246 _Reference_: 2247 OpenACC specification v2.0 (https://www.openacc.org), section 2248 3.2.22. 2249 2250 2251 File: libgomp.info, Node: acc_update_device, Next: acc_update_self, Prev: acc_delete, Up: OpenACC Runtime Library Routines 2252 2253 5.23 'acc_update_device' - Update device memory from mapped host memory. 2254 ======================================================================== 2255 2256 _Description_ 2257 This function updates the device copy from the previously mapped 2258 host memory. The host memory is specified with the host address A 2259 and a length of LEN bytes. 2260 2261 In Fortran, two (2) forms are supported. In the first form, A 2262 specifies a contiguous array section. The second form A specifies 2263 a variable or array element and LEN specifies the length in bytes. 2264 2265 _C/C++_: 2266 _Prototype_: 'acc_update_device(h_void *a, size_t len);' 2267 2268 _Fortran_: 2269 _Interface_: 'subroutine acc_update_device(a)' 2270 'type, dimension(:[,:]...) :: a' 2271 _Interface_: 'subroutine acc_update_device(a, len)' 2272 'type, dimension(:[,:]...) :: a' 2273 'integer len' 2274 2275 _Reference_: 2276 OpenACC specification v2.0 (https://www.openacc.org), section 2277 3.2.23. 2278 2279 2280 File: libgomp.info, Node: acc_update_self, Next: acc_map_data, Prev: acc_update_device, Up: OpenACC Runtime Library Routines 2281 2282 5.24 'acc_update_self' - Update host memory from mapped device memory. 2283 ====================================================================== 2284 2285 _Description_ 2286 This function updates the host copy from the previously mapped 2287 device memory. The host memory is specified with the host address 2288 A and a length of LEN bytes. 2289 2290 In Fortran, two (2) forms are supported. In the first form, A 2291 specifies a contiguous array section. The second form A specifies 2292 a variable or array element and LEN specifies the length in bytes. 2293 2294 _C/C++_: 2295 _Prototype_: 'acc_update_self(h_void *a, size_t len);' 2296 2297 _Fortran_: 2298 _Interface_: 'subroutine acc_update_self(a)' 2299 'type, dimension(:[,:]...) :: a' 2300 _Interface_: 'subroutine acc_update_self(a, len)' 2301 'type, dimension(:[,:]...) :: a' 2302 'integer len' 2303 2304 _Reference_: 2305 OpenACC specification v2.0 (https://www.openacc.org), section 2306 3.2.24. 2307 2308 2309 File: libgomp.info, Node: acc_map_data, Next: acc_unmap_data, Prev: acc_update_self, Up: OpenACC Runtime Library Routines 2310 2311 5.25 'acc_map_data' - Map previously allocated device memory to host memory. 2312 ============================================================================ 2313 2314 _Description_ 2315 This function maps previously allocated device and host memory. 2316 The device memory is specified with the device address D. The host 2317 memory is specified with the host address H and a length of LEN. 2318 2319 _C/C++_: 2320 _Prototype_: 'acc_map_data(h_void *h, d_void *d, size_t len);' 2321 2322 _Reference_: 2323 OpenACC specification v2.0 (https://www.openacc.org), section 2324 3.2.25. 2325 2326 2327 File: libgomp.info, Node: acc_unmap_data, Next: acc_deviceptr, Prev: acc_map_data, Up: OpenACC Runtime Library Routines 2328 2329 5.26 'acc_unmap_data' - Unmap device memory from host memory. 2330 ============================================================= 2331 2332 _Description_ 2333 This function unmaps previously mapped device and host memory. The 2334 latter specified by H. 2335 2336 _C/C++_: 2337 _Prototype_: 'acc_unmap_data(h_void *h);' 2338 2339 _Reference_: 2340 OpenACC specification v2.0 (https://www.openacc.org), section 2341 3.2.26. 2342 2343 2344 File: libgomp.info, Node: acc_deviceptr, Next: acc_hostptr, Prev: acc_unmap_data, Up: OpenACC Runtime Library Routines 2345 2346 5.27 'acc_deviceptr' - Get device pointer associated with specific host address. 2347 ================================================================================ 2348 2349 _Description_ 2350 This function returns the device address that has been mapped to 2351 the host address specified by H. 2352 2353 _C/C++_: 2354 _Prototype_: 'void *acc_deviceptr(h_void *h);' 2355 2356 _Reference_: 2357 OpenACC specification v2.0 (https://www.openacc.org), section 2358 3.2.27. 2359 2360 2361 File: libgomp.info, Node: acc_hostptr, Next: acc_is_present, Prev: acc_deviceptr, Up: OpenACC Runtime Library Routines 2362 2363 5.28 'acc_hostptr' - Get host pointer associated with specific device address. 2364 ============================================================================== 2365 2366 _Description_ 2367 This function returns the host address that has been mapped to the 2368 device address specified by D. 2369 2370 _C/C++_: 2371 _Prototype_: 'void *acc_hostptr(d_void *d);' 2372 2373 _Reference_: 2374 OpenACC specification v2.0 (https://www.openacc.org), section 2375 3.2.28. 2376 2377 2378 File: libgomp.info, Node: acc_is_present, Next: acc_memcpy_to_device, Prev: acc_hostptr, Up: OpenACC Runtime Library Routines 2379 2380 5.29 'acc_is_present' - Indicate whether host variable / array is present on device. 2381 ==================================================================================== 2382 2383 _Description_ 2384 This function indicates whether the specified host address in A and 2385 a length of LEN bytes is present on the device. In C/C++, a 2386 non-zero value is returned to indicate the presence of the mapped 2387 memory on the device. A zero is returned to indicate the memory is 2388 not mapped on the device. 2389 2390 In Fortran, two (2) forms are supported. In the first form, A 2391 specifies a contiguous array section. The second form A specifies 2392 a variable or array element and LEN specifies the length in bytes. 2393 If the host memory is mapped to device memory, then a 'true' is 2394 returned. Otherwise, a 'false' is return to indicate the mapped 2395 memory is not present. 2396 2397 _C/C++_: 2398 _Prototype_: 'int acc_is_present(h_void *a, size_t len);' 2399 2400 _Fortran_: 2401 _Interface_: 'function acc_is_present(a)' 2402 'type, dimension(:[,:]...) :: a' 2403 'logical acc_is_present' 2404 _Interface_: 'function acc_is_present(a, len)' 2405 'type, dimension(:[,:]...) :: a' 2406 'integer len' 2407 'logical acc_is_present' 2408 2409 _Reference_: 2410 OpenACC specification v2.0 (https://www.openacc.org), section 2411 3.2.29. 2412 2413 2414 File: libgomp.info, Node: acc_memcpy_to_device, Next: acc_memcpy_from_device, Prev: acc_is_present, Up: OpenACC Runtime Library Routines 2415 2416 5.30 'acc_memcpy_to_device' - Copy host memory to device memory. 2417 ================================================================ 2418 2419 _Description_ 2420 This function copies host memory specified by host address of SRC 2421 to device memory specified by the device address DEST for a length 2422 of BYTES bytes. 2423 2424 _C/C++_: 2425 _Prototype_: 'acc_memcpy_to_device(d_void *dest, h_void *src, size_t 2426 bytes);' 2427 2428 _Reference_: 2429 OpenACC specification v2.0 (https://www.openacc.org), section 2430 3.2.30. 2431 2432 2433 File: libgomp.info, Node: acc_memcpy_from_device, Next: acc_get_current_cuda_device, Prev: acc_memcpy_to_device, Up: OpenACC Runtime Library Routines 2434 2435 5.31 'acc_memcpy_from_device' - Copy device memory to host memory. 2436 ================================================================== 2437 2438 _Description_ 2439 This function copies host memory specified by host address of SRC 2440 from device memory specified by the device address DEST for a 2441 length of BYTES bytes. 2442 2443 _C/C++_: 2444 _Prototype_: 'acc_memcpy_from_device(d_void *dest, h_void *src, 2445 size_t bytes);' 2446 2447 _Reference_: 2448 OpenACC specification v2.0 (https://www.openacc.org), section 2449 3.2.31. 2450 2451 2452 File: libgomp.info, Node: acc_get_current_cuda_device, Next: acc_get_current_cuda_context, Prev: acc_memcpy_from_device, Up: OpenACC Runtime Library Routines 2453 2454 5.32 'acc_get_current_cuda_device' - Get CUDA device handle. 2455 ============================================================ 2456 2457 _Description_ 2458 This function returns the CUDA device handle. This handle is the 2459 same as used by the CUDA Runtime or Driver API's. 2460 2461 _C/C++_: 2462 _Prototype_: 'void *acc_get_current_cuda_device(void);' 2463 2464 _Reference_: 2465 OpenACC specification v2.0 (https://www.openacc.org), section 2466 A.2.1.1. 2467 2468 2469 File: libgomp.info, Node: acc_get_current_cuda_context, Next: acc_get_cuda_stream, Prev: acc_get_current_cuda_device, Up: OpenACC Runtime Library Routines 2470 2471 5.33 'acc_get_current_cuda_context' - Get CUDA context handle. 2472 ============================================================== 2473 2474 _Description_ 2475 This function returns the CUDA context handle. This handle is the 2476 same as used by the CUDA Runtime or Driver API's. 2477 2478 _C/C++_: 2479 _Prototype_: 'void *acc_get_current_cuda_context(void);' 2480 2481 _Reference_: 2482 OpenACC specification v2.0 (https://www.openacc.org), section 2483 A.2.1.2. 2484 2485 2486 File: libgomp.info, Node: acc_get_cuda_stream, Next: acc_set_cuda_stream, Prev: acc_get_current_cuda_context, Up: OpenACC Runtime Library Routines 2487 2488 5.34 'acc_get_cuda_stream' - Get CUDA stream handle. 2489 ==================================================== 2490 2491 _Description_ 2492 This function returns the CUDA stream handle for the queue ASYNC. 2493 This handle is the same as used by the CUDA Runtime or Driver 2494 API's. 2495 2496 _C/C++_: 2497 _Prototype_: 'void *acc_get_cuda_stream(int async);' 2498 2499 _Reference_: 2500 OpenACC specification v2.0 (https://www.openacc.org), section 2501 A.2.1.3. 2502 2503 2504 File: libgomp.info, Node: acc_set_cuda_stream, Prev: acc_get_cuda_stream, Up: OpenACC Runtime Library Routines 2505 2506 5.35 'acc_set_cuda_stream' - Set CUDA stream handle. 2507 ==================================================== 2508 2509 _Description_ 2510 This function associates the stream handle specified by STREAM with 2511 the queue ASYNC. 2512 2513 This cannot be used to change the stream handle associated with 2514 'acc_async_sync'. 2515 2516 The return value is not specified. 2517 2518 _C/C++_: 2519 _Prototype_: 'int acc_set_cuda_stream(int async, void *stream);' 2520 2521 _Reference_: 2522 OpenACC specification v2.0 (https://www.openacc.org), section 2523 A.2.1.4. 2524 2525 2526 File: libgomp.info, Node: OpenACC Environment Variables, Next: CUDA Streams Usage, Prev: OpenACC Runtime Library Routines, Up: Top 2527 2528 6 OpenACC Environment Variables 2529 ******************************* 2530 2531 The variables 'ACC_DEVICE_TYPE' and 'ACC_DEVICE_NUM' are defined by 2532 section 4 of the OpenACC specification in version 2.0. The variable 2533 'GCC_ACC_NOTIFY' is used for diagnostic purposes. 2534 2535 * Menu: 2536 2537 * ACC_DEVICE_TYPE:: 2538 * ACC_DEVICE_NUM:: 2539 * GCC_ACC_NOTIFY:: 2540 2541 2542 File: libgomp.info, Node: ACC_DEVICE_TYPE, Next: ACC_DEVICE_NUM, Up: OpenACC Environment Variables 2543 2544 6.1 'ACC_DEVICE_TYPE' 2545 ===================== 2546 2547 _Reference_: 2548 OpenACC specification v2.0 (https://www.openacc.org), section 4.1. 2549 2550 2551 File: libgomp.info, Node: ACC_DEVICE_NUM, Next: GCC_ACC_NOTIFY, Prev: ACC_DEVICE_TYPE, Up: OpenACC Environment Variables 2552 2553 6.2 'ACC_DEVICE_NUM' 2554 ==================== 2555 2556 _Reference_: 2557 OpenACC specification v2.0 (https://www.openacc.org), section 4.2. 2558 2559 2560 File: libgomp.info, Node: GCC_ACC_NOTIFY, Prev: ACC_DEVICE_NUM, Up: OpenACC Environment Variables 2561 2562 6.3 'GCC_ACC_NOTIFY' 2563 ==================== 2564 2565 _Description_: 2566 Print debug information pertaining to the accelerator. 2567 2568 2569 File: libgomp.info, Node: CUDA Streams Usage, Next: OpenACC Library Interoperability, Prev: OpenACC Environment Variables, Up: Top 2570 2571 7 CUDA Streams Usage 2572 ******************** 2573 2574 This applies to the 'nvptx' plugin only. 2575 2576 The library provides elements that perform asynchronous movement of 2577 data and asynchronous operation of computing constructs. This 2578 asynchronous functionality is implemented by making use of CUDA 2579 streams(1). 2580 2581 The primary means by that the asychronous functionality is accessed 2582 is through the use of those OpenACC directives which make use of the 2583 'async' and 'wait' clauses. When the 'async' clause is first used with 2584 a directive, it creates a CUDA stream. If an 'async-argument' is used 2585 with the 'async' clause, then the stream is associated with the 2586 specified 'async-argument'. 2587 2588 Following the creation of an association between a CUDA stream and 2589 the 'async-argument' of an 'async' clause, both the 'wait' clause and 2590 the 'wait' directive can be used. When either the clause or directive 2591 is used after stream creation, it creates a rendezvous point whereby 2592 execution waits until all operations associated with the 2593 'async-argument', that is, stream, have completed. 2594 2595 Normally, the management of the streams that are created as a result 2596 of using the 'async' clause, is done without any intervention by the 2597 caller. This implies the association between the 'async-argument' and 2598 the CUDA stream will be maintained for the lifetime of the program. 2599 However, this association can be changed through the use of the library 2600 function 'acc_set_cuda_stream'. When the function 'acc_set_cuda_stream' 2601 is called, the CUDA stream that was originally associated with the 2602 'async' clause will be destroyed. Caution should be taken when changing 2603 the association as subsequent references to the 'async-argument' refer 2604 to a different CUDA stream. 2605 2606 ---------- Footnotes ---------- 2607 2608 (1) See "Stream Management" in "CUDA Driver API", TRM-06703-001, 2609 Version 5.5, for additional information 2610 2611 2612 File: libgomp.info, Node: OpenACC Library Interoperability, Next: The libgomp ABI, Prev: CUDA Streams Usage, Up: Top 2613 2614 8 OpenACC Library Interoperability 2615 ********************************** 2616 2617 8.1 Introduction 2618 ================ 2619 2620 The OpenACC library uses the CUDA Driver API, and may interact with 2621 programs that use the Runtime library directly, or another library based 2622 on the Runtime library, e.g., CUBLAS(1). This chapter describes the use 2623 cases and what changes are required in order to use both the OpenACC 2624 library and the CUBLAS and Runtime libraries within a program. 2625 2626 8.2 First invocation: NVIDIA CUBLAS library API 2627 =============================================== 2628 2629 In this first use case (see below), a function in the CUBLAS library is 2630 called prior to any of the functions in the OpenACC library. More 2631 specifically, the function 'cublasCreate()'. 2632 2633 When invoked, the function initializes the library and allocates the 2634 hardware resources on the host and the device on behalf of the caller. 2635 Once the initialization and allocation has completed, a handle is 2636 returned to the caller. The OpenACC library also requires 2637 initialization and allocation of hardware resources. Since the CUBLAS 2638 library has already allocated the hardware resources for the device, all 2639 that is left to do is to initialize the OpenACC library and acquire the 2640 hardware resources on the host. 2641 2642 Prior to calling the OpenACC function that initializes the library 2643 and allocate the host hardware resources, you need to acquire the device 2644 number that was allocated during the call to 'cublasCreate()'. The 2645 invoking of the runtime library function 'cudaGetDevice()' accomplishes 2646 this. Once acquired, the device number is passed along with the device 2647 type as parameters to the OpenACC library function 2648 'acc_set_device_num()'. 2649 2650 Once the call to 'acc_set_device_num()' has completed, the OpenACC 2651 library uses the context that was created during the call to 2652 'cublasCreate()'. In other words, both libraries will be sharing the 2653 same context. 2654 2655 /* Create the handle */ 2656 s = cublasCreate(&h); 2657 if (s != CUBLAS_STATUS_SUCCESS) 2658 { 2659 fprintf(stderr, "cublasCreate failed %d\n", s); 2660 exit(EXIT_FAILURE); 2661 } 2662 2663 /* Get the device number */ 2664 e = cudaGetDevice(&dev); 2665 if (e != cudaSuccess) 2666 { 2667 fprintf(stderr, "cudaGetDevice failed %d\n", e); 2668 exit(EXIT_FAILURE); 2669 } 2670 2671 /* Initialize OpenACC library and use device 'dev' */ 2672 acc_set_device_num(dev, acc_device_nvidia); 2673 2674 Use Case 1 2675 2676 8.3 First invocation: OpenACC library API 2677 ========================================= 2678 2679 In this second use case (see below), a function in the OpenACC library 2680 is called prior to any of the functions in the CUBLAS library. More 2681 specificially, the function 'acc_set_device_num()'. 2682 2683 In the use case presented here, the function 'acc_set_device_num()' 2684 is used to both initialize the OpenACC library and allocate the hardware 2685 resources on the host and the device. In the call to the function, the 2686 call parameters specify which device to use and what device type to use, 2687 i.e., 'acc_device_nvidia'. It should be noted that this is but one 2688 method to initialize the OpenACC library and allocate the appropriate 2689 hardware resources. Other methods are available through the use of 2690 environment variables and these will be discussed in the next section. 2691 2692 Once the call to 'acc_set_device_num()' has completed, other OpenACC 2693 functions can be called as seen with multiple calls being made to 2694 'acc_copyin()'. In addition, calls can be made to functions in the 2695 CUBLAS library. In the use case a call to 'cublasCreate()' is made 2696 subsequent to the calls to 'acc_copyin()'. As seen in the previous use 2697 case, a call to 'cublasCreate()' initializes the CUBLAS library and 2698 allocates the hardware resources on the host and the device. However, 2699 since the device has already been allocated, 'cublasCreate()' will only 2700 initialize the CUBLAS library and allocate the appropriate hardware 2701 resources on the host. The context that was created as part of the 2702 OpenACC initialization is shared with the CUBLAS library, similarly to 2703 the first use case. 2704 2705 dev = 0; 2706 2707 acc_set_device_num(dev, acc_device_nvidia); 2708 2709 /* Copy the first set to the device */ 2710 d_X = acc_copyin(&h_X[0], N * sizeof (float)); 2711 if (d_X == NULL) 2712 { 2713 fprintf(stderr, "copyin error h_X\n"); 2714 exit(EXIT_FAILURE); 2715 } 2716 2717 /* Copy the second set to the device */ 2718 d_Y = acc_copyin(&h_Y1[0], N * sizeof (float)); 2719 if (d_Y == NULL) 2720 { 2721 fprintf(stderr, "copyin error h_Y1\n"); 2722 exit(EXIT_FAILURE); 2723 } 2724 2725 /* Create the handle */ 2726 s = cublasCreate(&h); 2727 if (s != CUBLAS_STATUS_SUCCESS) 2728 { 2729 fprintf(stderr, "cublasCreate failed %d\n", s); 2730 exit(EXIT_FAILURE); 2731 } 2732 2733 /* Perform saxpy using CUBLAS library function */ 2734 s = cublasSaxpy(h, N, &alpha, d_X, 1, d_Y, 1); 2735 if (s != CUBLAS_STATUS_SUCCESS) 2736 { 2737 fprintf(stderr, "cublasSaxpy failed %d\n", s); 2738 exit(EXIT_FAILURE); 2739 } 2740 2741 /* Copy the results from the device */ 2742 acc_memcpy_from_device(&h_Y1[0], d_Y, N * sizeof (float)); 2743 2744 Use Case 2 2745 2746 8.4 OpenACC library and environment variables 2747 ============================================= 2748 2749 There are two environment variables associated with the OpenACC library 2750 that may be used to control the device type and device number: 2751 'ACC_DEVICE_TYPE' and 'ACC_DEVICE_NUM', respecively. These two 2752 environement variables can be used as an alternative to calling 2753 'acc_set_device_num()'. As seen in the second use case, the device type 2754 and device number were specified using 'acc_set_device_num()'. If 2755 however, the aforementioned environment variables were set, then the 2756 call to 'acc_set_device_num()' would not be required. 2757 2758 The use of the environment variables is only relevant when an OpenACC 2759 function is called prior to a call to 'cudaCreate()'. If 'cudaCreate()' 2760 is called prior to a call to an OpenACC function, then you must call 2761 'acc_set_device_num()'(2) 2762 2763 ---------- Footnotes ---------- 2764 2765 (1) See section 2.26, "Interactions with the CUDA Driver API" in 2766 "CUDA Runtime API", Version 5.5, and section 2.27, "VDPAU 2767 Interoperability", in "CUDA Driver API", TRM-06703-001, Version 5.5, for 2768 additional information on library interoperability. 2769 2770 (2) More complete information about 'ACC_DEVICE_TYPE' and 2771 'ACC_DEVICE_NUM' can be found in sections 4.1 and 4.2 of the OpenACC 2772 (https://www.openacc.org) Application Programming Interface, Version 2773 2.0. 2774 2775 2776 File: libgomp.info, Node: The libgomp ABI, Next: Reporting Bugs, Prev: OpenACC Library Interoperability, Up: Top 2777 2778 9 The libgomp ABI 2779 ***************** 2780 2781 The following sections present notes on the external ABI as presented by 2782 libgomp. Only maintainers should need them. 2783 2784 * Menu: 2785 2786 * Implementing MASTER construct:: 2787 * Implementing CRITICAL construct:: 2788 * Implementing ATOMIC construct:: 2789 * Implementing FLUSH construct:: 2790 * Implementing BARRIER construct:: 2791 * Implementing THREADPRIVATE construct:: 2792 * Implementing PRIVATE clause:: 2793 * Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses:: 2794 * Implementing REDUCTION clause:: 2795 * Implementing PARALLEL construct:: 2796 * Implementing FOR construct:: 2797 * Implementing ORDERED construct:: 2798 * Implementing SECTIONS construct:: 2799 * Implementing SINGLE construct:: 2800 * Implementing OpenACC's PARALLEL construct:: 2801 2802 2803 File: libgomp.info, Node: Implementing MASTER construct, Next: Implementing CRITICAL construct, Up: The libgomp ABI 2804 2805 9.1 Implementing MASTER construct 2806 ================================= 2807 2808 if (omp_get_thread_num () == 0) 2809 block 2810 2811 Alternately, we generate two copies of the parallel subfunction and 2812 only include this in the version run by the master thread. Surely this 2813 is not worthwhile though... 2814 2815 2816 File: libgomp.info, Node: Implementing CRITICAL construct, Next: Implementing ATOMIC construct, Prev: Implementing MASTER construct, Up: The libgomp ABI 2817 2818 9.2 Implementing CRITICAL construct 2819 =================================== 2820 2821 Without a specified name, 2822 2823 void GOMP_critical_start (void); 2824 void GOMP_critical_end (void); 2825 2826 so that we don't get COPY relocations from libgomp to the main 2827 application. 2828 2829 With a specified name, use omp_set_lock and omp_unset_lock with name 2830 being transformed into a variable declared like 2831 2832 omp_lock_t gomp_critical_user_<name> __attribute__((common)) 2833 2834 Ideally the ABI would specify that all zero is a valid unlocked 2835 state, and so we wouldn't need to initialize this at startup. 2836 2837 2838 File: libgomp.info, Node: Implementing ATOMIC construct, Next: Implementing FLUSH construct, Prev: Implementing CRITICAL construct, Up: The libgomp ABI 2839 2840 9.3 Implementing ATOMIC construct 2841 ================================= 2842 2843 The target should implement the '__sync' builtins. 2844 2845 Failing that we could add 2846 2847 void GOMP_atomic_enter (void) 2848 void GOMP_atomic_exit (void) 2849 2850 which reuses the regular lock code, but with yet another lock object 2851 private to the library. 2852 2853 2854 File: libgomp.info, Node: Implementing FLUSH construct, Next: Implementing BARRIER construct, Prev: Implementing ATOMIC construct, Up: The libgomp ABI 2855 2856 9.4 Implementing FLUSH construct 2857 ================================ 2858 2859 Expands to the '__sync_synchronize' builtin. 2860 2861 2862 File: libgomp.info, Node: Implementing BARRIER construct, Next: Implementing THREADPRIVATE construct, Prev: Implementing FLUSH construct, Up: The libgomp ABI 2863 2864 9.5 Implementing BARRIER construct 2865 ================================== 2866 2867 void GOMP_barrier (void) 2868 2869 2870 File: libgomp.info, Node: Implementing THREADPRIVATE construct, Next: Implementing PRIVATE clause, Prev: Implementing BARRIER construct, Up: The libgomp ABI 2871 2872 9.6 Implementing THREADPRIVATE construct 2873 ======================================== 2874 2875 In _most_ cases we can map this directly to '__thread'. Except that OMP 2876 allows constructors for C++ objects. We can either refuse to support 2877 this (how often is it used?) or we can implement something akin to 2878 .ctors. 2879 2880 Even more ideally, this ctor feature is handled by extensions to the 2881 main pthreads library. Failing that, we can have a set of entry points 2882 to register ctor functions to be called. 2883 2884 2885 File: libgomp.info, Node: Implementing PRIVATE clause, Next: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses, Prev: Implementing THREADPRIVATE construct, Up: The libgomp ABI 2886 2887 9.7 Implementing PRIVATE clause 2888 =============================== 2889 2890 In association with a PARALLEL, or within the lexical extent of a 2891 PARALLEL block, the variable becomes a local variable in the parallel 2892 subfunction. 2893 2894 In association with FOR or SECTIONS blocks, create a new automatic 2895 variable within the current function. This preserves the semantic of 2896 new variable creation. 2897 2898 2899 File: libgomp.info, Node: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses, Next: Implementing REDUCTION clause, Prev: Implementing PRIVATE clause, Up: The libgomp ABI 2900 2901 9.8 Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses 2902 ======================================================================== 2903 2904 This seems simple enough for PARALLEL blocks. Create a private struct 2905 for communicating between the parent and subfunction. In the parent, 2906 copy in values for scalar and "small" structs; copy in addresses for 2907 others TREE_ADDRESSABLE types. In the subfunction, copy the value into 2908 the local variable. 2909 2910 It is not clear what to do with bare FOR or SECTION blocks. The only 2911 thing I can figure is that we do something like: 2912 2913 #pragma omp for firstprivate(x) lastprivate(y) 2914 for (int i = 0; i < n; ++i) 2915 body; 2916 2917 which becomes 2918 2919 { 2920 int x = x, y; 2921 2922 // for stuff 2923 2924 if (i == n) 2925 y = y; 2926 } 2927 2928 where the "x=x" and "y=y" assignments actually have different uids 2929 for the two variables, i.e. not something you could write directly in 2930 C. Presumably this only makes sense if the "outer" x and y are global 2931 variables. 2932 2933 COPYPRIVATE would work the same way, except the structure broadcast 2934 would have to happen via SINGLE machinery instead. 2935 2936 2937 File: libgomp.info, Node: Implementing REDUCTION clause, Next: Implementing PARALLEL construct, Prev: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses, Up: The libgomp ABI 2938 2939 9.9 Implementing REDUCTION clause 2940 ================================= 2941 2942 The private struct mentioned in the previous section should have a 2943 pointer to an array of the type of the variable, indexed by the thread's 2944 TEAM_ID. The thread stores its final value into the array, and after 2945 the barrier, the master thread iterates over the array to collect the 2946 values. 2947 2948 2949 File: libgomp.info, Node: Implementing PARALLEL construct, Next: Implementing FOR construct, Prev: Implementing REDUCTION clause, Up: The libgomp ABI 2950 2951 9.10 Implementing PARALLEL construct 2952 ==================================== 2953 2954 #pragma omp parallel 2955 { 2956 body; 2957 } 2958 2959 becomes 2960 2961 void subfunction (void *data) 2962 { 2963 use data; 2964 body; 2965 } 2966 2967 setup data; 2968 GOMP_parallel_start (subfunction, &data, num_threads); 2969 subfunction (&data); 2970 GOMP_parallel_end (); 2971 2972 void GOMP_parallel_start (void (*fn)(void *), void *data, unsigned num_threads) 2973 2974 The FN argument is the subfunction to be run in parallel. 2975 2976 The DATA argument is a pointer to a structure used to communicate 2977 data in and out of the subfunction, as discussed above with respect to 2978 FIRSTPRIVATE et al. 2979 2980 The NUM_THREADS argument is 1 if an IF clause is present and false, 2981 or the value of the NUM_THREADS clause, if present, or 0. 2982 2983 The function needs to create the appropriate number of threads and/or 2984 launch them from the dock. It needs to create the team structure and 2985 assign team ids. 2986 2987 void GOMP_parallel_end (void) 2988 2989 Tears down the team and returns us to the previous 2990 'omp_in_parallel()' state. 2991 2992 2993 File: libgomp.info, Node: Implementing FOR construct, Next: Implementing ORDERED construct, Prev: Implementing PARALLEL construct, Up: The libgomp ABI 2994 2995 9.11 Implementing FOR construct 2996 =============================== 2997 2998 #pragma omp parallel for 2999 for (i = lb; i <= ub; i++) 3000 body; 3001 3002 becomes 3003 3004 void subfunction (void *data) 3005 { 3006 long _s0, _e0; 3007 while (GOMP_loop_static_next (&_s0, &_e0)) 3008 { 3009 long _e1 = _e0, i; 3010 for (i = _s0; i < _e1; i++) 3011 body; 3012 } 3013 GOMP_loop_end_nowait (); 3014 } 3015 3016 GOMP_parallel_loop_static (subfunction, NULL, 0, lb, ub+1, 1, 0); 3017 subfunction (NULL); 3018 GOMP_parallel_end (); 3019 3020 #pragma omp for schedule(runtime) 3021 for (i = 0; i < n; i++) 3022 body; 3023 3024 becomes 3025 3026 { 3027 long i, _s0, _e0; 3028 if (GOMP_loop_runtime_start (0, n, 1, &_s0, &_e0)) 3029 do { 3030 long _e1 = _e0; 3031 for (i = _s0, i < _e0; i++) 3032 body; 3033 } while (GOMP_loop_runtime_next (&_s0, _&e0)); 3034 GOMP_loop_end (); 3035 } 3036 3037 Note that while it looks like there is trickiness to propagating a 3038 non-constant STEP, there isn't really. We're explicitly allowed to 3039 evaluate it as many times as we want, and any variables involved should 3040 automatically be handled as PRIVATE or SHARED like any other variables. 3041 So the expression should remain evaluable in the subfunction. We can 3042 also pull it into a local variable if we like, but since its supposed to 3043 remain unchanged, we can also not if we like. 3044 3045 If we have SCHEDULE(STATIC), and no ORDERED, then we ought to be able 3046 to get away with no work-sharing context at all, since we can simply 3047 perform the arithmetic directly in each thread to divide up the 3048 iterations. Which would mean that we wouldn't need to call any of these 3049 routines. 3050 3051 There are separate routines for handling loops with an ORDERED 3052 clause. Bookkeeping for that is non-trivial... 3053 3054 3055 File: libgomp.info, Node: Implementing ORDERED construct, Next: Implementing SECTIONS construct, Prev: Implementing FOR construct, Up: The libgomp ABI 3056 3057 9.12 Implementing ORDERED construct 3058 =================================== 3059 3060 void GOMP_ordered_start (void) 3061 void GOMP_ordered_end (void) 3062 3063 3064 File: libgomp.info, Node: Implementing SECTIONS construct, Next: Implementing SINGLE construct, Prev: Implementing ORDERED construct, Up: The libgomp ABI 3065 3066 9.13 Implementing SECTIONS construct 3067 ==================================== 3068 3069 A block as 3070 3071 #pragma omp sections 3072 { 3073 #pragma omp section 3074 stmt1; 3075 #pragma omp section 3076 stmt2; 3077 #pragma omp section 3078 stmt3; 3079 } 3080 3081 becomes 3082 3083 for (i = GOMP_sections_start (3); i != 0; i = GOMP_sections_next ()) 3084 switch (i) 3085 { 3086 case 1: 3087 stmt1; 3088 break; 3089 case 2: 3090 stmt2; 3091 break; 3092 case 3: 3093 stmt3; 3094 break; 3095 } 3096 GOMP_barrier (); 3097 3098 3099 File: libgomp.info, Node: Implementing SINGLE construct, Next: Implementing OpenACC's PARALLEL construct, Prev: Implementing SECTIONS construct, Up: The libgomp ABI 3100 3101 9.14 Implementing SINGLE construct 3102 ================================== 3103 3104 A block like 3105 3106 #pragma omp single 3107 { 3108 body; 3109 } 3110 3111 becomes 3112 3113 if (GOMP_single_start ()) 3114 body; 3115 GOMP_barrier (); 3116 3117 while 3118 3119 #pragma omp single copyprivate(x) 3120 body; 3121 3122 becomes 3123 3124 datap = GOMP_single_copy_start (); 3125 if (datap == NULL) 3126 { 3127 body; 3128 data.x = x; 3129 GOMP_single_copy_end (&data); 3130 } 3131 else 3132 x = datap->x; 3133 GOMP_barrier (); 3134 3135 3136 File: libgomp.info, Node: Implementing OpenACC's PARALLEL construct, Prev: Implementing SINGLE construct, Up: The libgomp ABI 3137 3138 9.15 Implementing OpenACC's PARALLEL construct 3139 ============================================== 3140 3141 void GOACC_parallel () 3142 3143 3144 File: libgomp.info, Node: Reporting Bugs, Next: Copying, Prev: The libgomp ABI, Up: Top 3145 3146 10 Reporting Bugs 3147 ***************** 3148 3149 Bugs in the GNU Offloading and Multi Processing Runtime Library should 3150 be reported via Bugzilla (http://gcc.gnu.org/bugzilla/). Please add 3151 "openacc", or "openmp", or both to the keywords field in the bug report, 3152 as appropriate. 3153 3154 3155 File: libgomp.info, Node: Copying, Next: GNU Free Documentation License, Prev: Reporting Bugs, Up: Top 3156 3157 GNU General Public License 3158 ************************** 3159 3160 Version 3, 29 June 2007 3161 3162 Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/> 3163 3164 Everyone is permitted to copy and distribute verbatim copies of this 3165 license document, but changing it is not allowed. 3166 3167 Preamble 3168 ======== 3169 3170 The GNU General Public License is a free, copyleft license for software 3171 and other kinds of works. 3172 3173 The licenses for most software and other practical works are designed 3174 to take away your freedom to share and change the works. By contrast, 3175 the GNU General Public License is intended to guarantee your freedom to 3176 share and change all versions of a program-to make sure it remains free 3177 software for all its users. We, the Free Software Foundation, use the 3178 GNU General Public License for most of our software; it applies also to 3179 any other work released this way by its authors. You can apply it to 3180 your programs, too. 3181 3182 When we speak of free software, we are referring to freedom, not 3183 price. Our General Public Licenses are designed to make sure that you 3184 have the freedom to distribute copies of free software (and charge for 3185 them if you wish), that you receive source code or can get it if you 3186 want it, that you can change the software or use pieces of it in new 3187 free programs, and that you know you can do these things. 3188 3189 To protect your rights, we need to prevent others from denying you 3190 these rights or asking you to surrender the rights. Therefore, you have 3191 certain responsibilities if you distribute copies of the software, or if 3192 you modify it: responsibilities to respect the freedom of others. 3193 3194 For example, if you distribute copies of such a program, whether 3195 gratis or for a fee, you must pass on to the recipients the same 3196 freedoms that you received. You must make sure that they, too, receive 3197 or can get the source code. And you must show them these terms so they 3198 know their rights. 3199 3200 Developers that use the GNU GPL protect your rights with two steps: 3201 (1) assert copyright on the software, and (2) offer you this License 3202 giving you legal permission to copy, distribute and/or modify it. 3203 3204 For the developers' and authors' protection, the GPL clearly explains 3205 that there is no warranty for this free software. For both users' and 3206 authors' sake, the GPL requires that modified versions be marked as 3207 changed, so that their problems will not be attributed erroneously to 3208 authors of previous versions. 3209 3210 Some devices are designed to deny users access to install or run 3211 modified versions of the software inside them, although the manufacturer 3212 can do so. This is fundamentally incompatible with the aim of 3213 protecting users' freedom to change the software. The systematic 3214 pattern of such abuse occurs in the area of products for individuals to 3215 use, which is precisely where it is most unacceptable. Therefore, we 3216 have designed this version of the GPL to prohibit the practice for those 3217 products. If such problems arise substantially in other domains, we 3218 stand ready to extend this provision to those domains in future versions 3219 of the GPL, as needed to protect the freedom of users. 3220 3221 Finally, every program is threatened constantly by software patents. 3222 States should not allow patents to restrict development and use of 3223 software on general-purpose computers, but in those that do, we wish to 3224 avoid the special danger that patents applied to a free program could 3225 make it effectively proprietary. To prevent this, the GPL assures that 3226 patents cannot be used to render the program non-free. 3227 3228 The precise terms and conditions for copying, distribution and 3229 modification follow. 3230 3231 TERMS AND CONDITIONS 3232 ==================== 3233 3234 0. Definitions. 3235 3236 "This License" refers to version 3 of the GNU General Public 3237 License. 3238 3239 "Copyright" also means copyright-like laws that apply to other 3240 kinds of works, such as semiconductor masks. 3241 3242 "The Program" refers to any copyrightable work licensed under this 3243 License. Each licensee is addressed as "you". "Licensees" and 3244 "recipients" may be individuals or organizations. 3245 3246 To "modify" a work means to copy from or adapt all or part of the 3247 work in a fashion requiring copyright permission, other than the 3248 making of an exact copy. The resulting work is called a "modified 3249 version" of the earlier work or a work "based on" the earlier work. 3250 3251 A "covered work" means either the unmodified Program or a work 3252 based on the Program. 3253 3254 To "propagate" a work means to do anything with it that, without 3255 permission, would make you directly or secondarily liable for 3256 infringement under applicable copyright law, except executing it on 3257 a computer or modifying a private copy. Propagation includes 3258 copying, distribution (with or without modification), making 3259 available to the public, and in some countries other activities as 3260 well. 3261 3262 To "convey" a work means any kind of propagation that enables other 3263 parties to make or receive copies. Mere interaction with a user 3264 through a computer network, with no transfer of a copy, is not 3265 conveying. 3266 3267 An interactive user interface displays "Appropriate Legal Notices" 3268 to the extent that it includes a convenient and prominently visible 3269 feature that (1) displays an appropriate copyright notice, and (2) 3270 tells the user that there is no warranty for the work (except to 3271 the extent that warranties are provided), that licensees may convey 3272 the work under this License, and how to view a copy of this 3273 License. If the interface presents a list of user commands or 3274 options, such as a menu, a prominent item in the list meets this 3275 criterion. 3276 3277 1. Source Code. 3278 3279 The "source code" for a work means the preferred form of the work 3280 for making modifications to it. "Object code" means any non-source 3281 form of a work. 3282 3283 A "Standard Interface" means an interface that either is an 3284 official standard defined by a recognized standards body, or, in 3285 the case of interfaces specified for a particular programming 3286 language, one that is widely used among developers working in that 3287 language. 3288 3289 The "System Libraries" of an executable work include anything, 3290 other than the work as a whole, that (a) is included in the normal 3291 form of packaging a Major Component, but which is not part of that 3292 Major Component, and (b) serves only to enable use of the work with 3293 that Major Component, or to implement a Standard Interface for 3294 which an implementation is available to the public in source code 3295 form. A "Major Component", in this context, means a major 3296 essential component (kernel, window system, and so on) of the 3297 specific operating system (if any) on which the executable work 3298 runs, or a compiler used to produce the work, or an object code 3299 interpreter used to run it. 3300 3301 The "Corresponding Source" for a work in object code form means all 3302 the source code needed to generate, install, and (for an executable 3303 work) run the object code and to modify the work, including scripts 3304 to control those activities. However, it does not include the 3305 work's System Libraries, or general-purpose tools or generally 3306 available free programs which are used unmodified in performing 3307 those activities but which are not part of the work. For example, 3308 Corresponding Source includes interface definition files associated 3309 with source files for the work, and the source code for shared 3310 libraries and dynamically linked subprograms that the work is 3311 specifically designed to require, such as by intimate data 3312 communication or control flow between those subprograms and other 3313 parts of the work. 3314 3315 The Corresponding Source need not include anything that users can 3316 regenerate automatically from other parts of the Corresponding 3317 Source. 3318 3319 The Corresponding Source for a work in source code form is that 3320 same work. 3321 3322 2. Basic Permissions. 3323 3324 All rights granted under this License are granted for the term of 3325 copyright on the Program, and are irrevocable provided the stated 3326 conditions are met. This License explicitly affirms your unlimited 3327 permission to run the unmodified Program. The output from running 3328 a covered work is covered by this License only if the output, given 3329 its content, constitutes a covered work. This License acknowledges 3330 your rights of fair use or other equivalent, as provided by 3331 copyright law. 3332 3333 You may make, run and propagate covered works that you do not 3334 convey, without conditions so long as your license otherwise 3335 remains in force. You may convey covered works to others for the 3336 sole purpose of having them make modifications exclusively for you, 3337 or provide you with facilities for running those works, provided 3338 that you comply with the terms of this License in conveying all 3339 material for which you do not control copyright. Those thus making 3340 or running the covered works for you must do so exclusively on your 3341 behalf, under your direction and control, on terms that prohibit 3342 them from making any copies of your copyrighted material outside 3343 their relationship with you. 3344 3345 Conveying under any other circumstances is permitted solely under 3346 the conditions stated below. Sublicensing is not allowed; section 3347 10 makes it unnecessary. 3348 3349 3. Protecting Users' Legal Rights From Anti-Circumvention Law. 3350 3351 No covered work shall be deemed part of an effective technological 3352 measure under any applicable law fulfilling obligations under 3353 article 11 of the WIPO copyright treaty adopted on 20 December 3354 1996, or similar laws prohibiting or restricting circumvention of 3355 such measures. 3356 3357 When you convey a covered work, you waive any legal power to forbid 3358 circumvention of technological measures to the extent such 3359 circumvention is effected by exercising rights under this License 3360 with respect to the covered work, and you disclaim any intention to 3361 limit operation or modification of the work as a means of 3362 enforcing, against the work's users, your or third parties' legal 3363 rights to forbid circumvention of technological measures. 3364 3365 4. Conveying Verbatim Copies. 3366 3367 You may convey verbatim copies of the Program's source code as you 3368 receive it, in any medium, provided that you conspicuously and 3369 appropriately publish on each copy an appropriate copyright notice; 3370 keep intact all notices stating that this License and any 3371 non-permissive terms added in accord with section 7 apply to the 3372 code; keep intact all notices of the absence of any warranty; and 3373 give all recipients a copy of this License along with the Program. 3374 3375 You may charge any price or no price for each copy that you convey, 3376 and you may offer support or warranty protection for a fee. 3377 3378 5. Conveying Modified Source Versions. 3379 3380 You may convey a work based on the Program, or the modifications to 3381 produce it from the Program, in the form of source code under the 3382 terms of section 4, provided that you also meet all of these 3383 conditions: 3384 3385 a. The work must carry prominent notices stating that you 3386 modified it, and giving a relevant date. 3387 3388 b. The work must carry prominent notices stating that it is 3389 released under this License and any conditions added under 3390 section 7. This requirement modifies the requirement in 3391 section 4 to "keep intact all notices". 3392 3393 c. You must license the entire work, as a whole, under this 3394 License to anyone who comes into possession of a copy. This 3395 License will therefore apply, along with any applicable 3396 section 7 additional terms, to the whole of the work, and all 3397 its parts, regardless of how they are packaged. This License 3398 gives no permission to license the work in any other way, but 3399 it does not invalidate such permission if you have separately 3400 received it. 3401 3402 d. If the work has interactive user interfaces, each must display 3403 Appropriate Legal Notices; however, if the Program has 3404 interactive interfaces that do not display Appropriate Legal 3405 Notices, your work need not make them do so. 3406 3407 A compilation of a covered work with other separate and independent 3408 works, which are not by their nature extensions of the covered 3409 work, and which are not combined with it such as to form a larger 3410 program, in or on a volume of a storage or distribution medium, is 3411 called an "aggregate" if the compilation and its resulting 3412 copyright are not used to limit the access or legal rights of the 3413 compilation's users beyond what the individual works permit. 3414 Inclusion of a covered work in an aggregate does not cause this 3415 License to apply to the other parts of the aggregate. 3416 3417 6. Conveying Non-Source Forms. 3418 3419 You may convey a covered work in object code form under the terms 3420 of sections 4 and 5, provided that you also convey the 3421 machine-readable Corresponding Source under the terms of this 3422 License, in one of these ways: 3423 3424 a. Convey the object code in, or embodied in, a physical product 3425 (including a physical distribution medium), accompanied by the 3426 Corresponding Source fixed on a durable physical medium 3427 customarily used for software interchange. 3428 3429 b. Convey the object code in, or embodied in, a physical product 3430 (including a physical distribution medium), accompanied by a 3431 written offer, valid for at least three years and valid for as 3432 long as you offer spare parts or customer support for that 3433 product model, to give anyone who possesses the object code 3434 either (1) a copy of the Corresponding Source for all the 3435 software in the product that is covered by this License, on a 3436 durable physical medium customarily used for software 3437 interchange, for a price no more than your reasonable cost of 3438 physically performing this conveying of source, or (2) access 3439 to copy the Corresponding Source from a network server at no 3440 charge. 3441 3442 c. Convey individual copies of the object code with a copy of the 3443 written offer to provide the Corresponding Source. This 3444 alternative is allowed only occasionally and noncommercially, 3445 and only if you received the object code with such an offer, 3446 in accord with subsection 6b. 3447 3448 d. Convey the object code by offering access from a designated 3449 place (gratis or for a charge), and offer equivalent access to 3450 the Corresponding Source in the same way through the same 3451 place at no further charge. You need not require recipients 3452 to copy the Corresponding Source along with the object code. 3453 If the place to copy the object code is a network server, the 3454 Corresponding Source may be on a different server (operated by 3455 you or a third party) that supports equivalent copying 3456 facilities, provided you maintain clear directions next to the 3457 object code saying where to find the Corresponding Source. 3458 Regardless of what server hosts the Corresponding Source, you 3459 remain obligated to ensure that it is available for as long as 3460 needed to satisfy these requirements. 3461 3462 e. Convey the object code using peer-to-peer transmission, 3463 provided you inform other peers where the object code and 3464 Corresponding Source of the work are being offered to the 3465 general public at no charge under subsection 6d. 3466 3467 A separable portion of the object code, whose source code is 3468 excluded from the Corresponding Source as a System Library, need 3469 not be included in conveying the object code work. 3470 3471 A "User Product" is either (1) a "consumer product", which means 3472 any tangible personal property which is normally used for personal, 3473 family, or household purposes, or (2) anything designed or sold for 3474 incorporation into a dwelling. In determining whether a product is 3475 a consumer product, doubtful cases shall be resolved in favor of 3476 coverage. For a particular product received by a particular user, 3477 "normally used" refers to a typical or common use of that class of 3478 product, regardless of the status of the particular user or of the 3479 way in which the particular user actually uses, or expects or is 3480 expected to use, the product. A product is a consumer product 3481 regardless of whether the product has substantial commercial, 3482 industrial or non-consumer uses, unless such uses represent the 3483 only significant mode of use of the product. 3484 3485 "Installation Information" for a User Product means any methods, 3486 procedures, authorization keys, or other information required to 3487 install and execute modified versions of a covered work in that 3488 User Product from a modified version of its Corresponding Source. 3489 The information must suffice to ensure that the continued 3490 functioning of the modified object code is in no case prevented or 3491 interfered with solely because modification has been made. 3492 3493 If you convey an object code work under this section in, or with, 3494 or specifically for use in, a User Product, and the conveying 3495 occurs as part of a transaction in which the right of possession 3496 and use of the User Product is transferred to the recipient in 3497 perpetuity or for a fixed term (regardless of how the transaction 3498 is characterized), the Corresponding Source conveyed under this 3499 section must be accompanied by the Installation Information. But 3500 this requirement does not apply if neither you nor any third party 3501 retains the ability to install modified object code on the User 3502 Product (for example, the work has been installed in ROM). 3503 3504 The requirement to provide Installation Information does not 3505 include a requirement to continue to provide support service, 3506 warranty, or updates for a work that has been modified or installed 3507 by the recipient, or for the User Product in which it has been 3508 modified or installed. Access to a network may be denied when the 3509 modification itself materially and adversely affects the operation 3510 of the network or violates the rules and protocols for 3511 communication across the network. 3512 3513 Corresponding Source conveyed, and Installation Information 3514 provided, in accord with this section must be in a format that is 3515 publicly documented (and with an implementation available to the 3516 public in source code form), and must require no special password 3517 or key for unpacking, reading or copying. 3518 3519 7. Additional Terms. 3520 3521 "Additional permissions" are terms that supplement the terms of 3522 this License by making exceptions from one or more of its 3523 conditions. Additional permissions that are applicable to the 3524 entire Program shall be treated as though they were included in 3525 this License, to the extent that they are valid under applicable 3526 law. If additional permissions apply only to part of the Program, 3527 that part may be used separately under those permissions, but the 3528 entire Program remains governed by this License without regard to 3529 the additional permissions. 3530 3531 When you convey a copy of a covered work, you may at your option 3532 remove any additional permissions from that copy, or from any part 3533 of it. (Additional permissions may be written to require their own 3534 removal in certain cases when you modify the work.) You may place 3535 additional permissions on material, added by you to a covered work, 3536 for which you have or can give appropriate copyright permission. 3537 3538 Notwithstanding any other provision of this License, for material 3539 you add to a covered work, you may (if authorized by the copyright 3540 holders of that material) supplement the terms of this License with 3541 terms: 3542 3543 a. Disclaiming warranty or limiting liability differently from 3544 the terms of sections 15 and 16 of this License; or 3545 3546 b. Requiring preservation of specified reasonable legal notices 3547 or author attributions in that material or in the Appropriate 3548 Legal Notices displayed by works containing it; or 3549 3550 c. Prohibiting misrepresentation of the origin of that material, 3551 or requiring that modified versions of such material be marked 3552 in reasonable ways as different from the original version; or 3553 3554 d. Limiting the use for publicity purposes of names of licensors 3555 or authors of the material; or 3556 3557 e. Declining to grant rights under trademark law for use of some 3558 trade names, trademarks, or service marks; or 3559 3560 f. Requiring indemnification of licensors and authors of that 3561 material by anyone who conveys the material (or modified 3562 versions of it) with contractual assumptions of liability to 3563 the recipient, for any liability that these contractual 3564 assumptions directly impose on those licensors and authors. 3565 3566 All other non-permissive additional terms are considered "further 3567 restrictions" within the meaning of section 10. If the Program as 3568 you received it, or any part of it, contains a notice stating that 3569 it is governed by this License along with a term that is a further 3570 restriction, you may remove that term. If a license document 3571 contains a further restriction but permits relicensing or conveying 3572 under this License, you may add to a covered work material governed 3573 by the terms of that license document, provided that the further 3574 restriction does not survive such relicensing or conveying. 3575 3576 If you add terms to a covered work in accord with this section, you 3577 must place, in the relevant source files, a statement of the 3578 additional terms that apply to those files, or a notice indicating 3579 where to find the applicable terms. 3580 3581 Additional terms, permissive or non-permissive, may be stated in 3582 the form of a separately written license, or stated as exceptions; 3583 the above requirements apply either way. 3584 3585 8. Termination. 3586 3587 You may not propagate or modify a covered work except as expressly 3588 provided under this License. Any attempt otherwise to propagate or 3589 modify it is void, and will automatically terminate your rights 3590 under this License (including any patent licenses granted under the 3591 third paragraph of section 11). 3592 3593 However, if you cease all violation of this License, then your 3594 license from a particular copyright holder is reinstated (a) 3595 provisionally, unless and until the copyright holder explicitly and 3596 finally terminates your license, and (b) permanently, if the 3597 copyright holder fails to notify you of the violation by some 3598 reasonable means prior to 60 days after the cessation. 3599 3600 Moreover, your license from a particular copyright holder is 3601 reinstated permanently if the copyright holder notifies you of the 3602 violation by some reasonable means, this is the first time you have 3603 received notice of violation of this License (for any work) from 3604 that copyright holder, and you cure the violation prior to 30 days 3605 after your receipt of the notice. 3606 3607 Termination of your rights under this section does not terminate 3608 the licenses of parties who have received copies or rights from you 3609 under this License. If your rights have been terminated and not 3610 permanently reinstated, you do not qualify to receive new licenses 3611 for the same material under section 10. 3612 3613 9. Acceptance Not Required for Having Copies. 3614 3615 You are not required to accept this License in order to receive or 3616 run a copy of the Program. Ancillary propagation of a covered work 3617 occurring solely as a consequence of using peer-to-peer 3618 transmission to receive a copy likewise does not require 3619 acceptance. However, nothing other than this License grants you 3620 permission to propagate or modify any covered work. These actions 3621 infringe copyright if you do not accept this License. Therefore, 3622 by modifying or propagating a covered work, you indicate your 3623 acceptance of this License to do so. 3624 3625 10. Automatic Licensing of Downstream Recipients. 3626 3627 Each time you convey a covered work, the recipient automatically 3628 receives a license from the original licensors, to run, modify and 3629 propagate that work, subject to this License. You are not 3630 responsible for enforcing compliance by third parties with this 3631 License. 3632 3633 An "entity transaction" is a transaction transferring control of an 3634 organization, or substantially all assets of one, or subdividing an 3635 organization, or merging organizations. If propagation of a 3636 covered work results from an entity transaction, each party to that 3637 transaction who receives a copy of the work also receives whatever 3638 licenses to the work the party's predecessor in interest had or 3639 could give under the previous paragraph, plus a right to possession 3640 of the Corresponding Source of the work from the predecessor in 3641 interest, if the predecessor has it or can get it with reasonable 3642 efforts. 3643 3644 You may not impose any further restrictions on the exercise of the 3645 rights granted or affirmed under this License. For example, you 3646 may not impose a license fee, royalty, or other charge for exercise 3647 of rights granted under this License, and you may not initiate 3648 litigation (including a cross-claim or counterclaim in a lawsuit) 3649 alleging that any patent claim is infringed by making, using, 3650 selling, offering for sale, or importing the Program or any portion 3651 of it. 3652 3653 11. Patents. 3654 3655 A "contributor" is a copyright holder who authorizes use under this 3656 License of the Program or a work on which the Program is based. 3657 The work thus licensed is called the contributor's "contributor 3658 version". 3659 3660 A contributor's "essential patent claims" are all patent claims 3661 owned or controlled by the contributor, whether already acquired or 3662 hereafter acquired, that would be infringed by some manner, 3663 permitted by this License, of making, using, or selling its 3664 contributor version, but do not include claims that would be 3665 infringed only as a consequence of further modification of the 3666 contributor version. For purposes of this definition, "control" 3667 includes the right to grant patent sublicenses in a manner 3668 consistent with the requirements of this License. 3669 3670 Each contributor grants you a non-exclusive, worldwide, 3671 royalty-free patent license under the contributor's essential 3672 patent claims, to make, use, sell, offer for sale, import and 3673 otherwise run, modify and propagate the contents of its contributor 3674 version. 3675 3676 In the following three paragraphs, a "patent license" is any 3677 express agreement or commitment, however denominated, not to 3678 enforce a patent (such as an express permission to practice a 3679 patent or covenant not to sue for patent infringement). To "grant" 3680 such a patent license to a party means to make such an agreement or 3681 commitment not to enforce a patent against the party. 3682 3683 If you convey a covered work, knowingly relying on a patent 3684 license, and the Corresponding Source of the work is not available 3685 for anyone to copy, free of charge and under the terms of this 3686 License, through a publicly available network server or other 3687 readily accessible means, then you must either (1) cause the 3688 Corresponding Source to be so available, or (2) arrange to deprive 3689 yourself of the benefit of the patent license for this particular 3690 work, or (3) arrange, in a manner consistent with the requirements 3691 of this License, to extend the patent license to downstream 3692 recipients. "Knowingly relying" means you have actual knowledge 3693 that, but for the patent license, your conveying the covered work 3694 in a country, or your recipient's use of the covered work in a 3695 country, would infringe one or more identifiable patents in that 3696 country that you have reason to believe are valid. 3697 3698 If, pursuant to or in connection with a single transaction or 3699 arrangement, you convey, or propagate by procuring conveyance of, a 3700 covered work, and grant a patent license to some of the parties 3701 receiving the covered work authorizing them to use, propagate, 3702 modify or convey a specific copy of the covered work, then the 3703 patent license you grant is automatically extended to all 3704 recipients of the covered work and works based on it. 3705 3706 A patent license is "discriminatory" if it does not include within 3707 the scope of its coverage, prohibits the exercise of, or is 3708 conditioned on the non-exercise of one or more of the rights that 3709 are specifically granted under this License. You may not convey a 3710 covered work if you are a party to an arrangement with a third 3711 party that is in the business of distributing software, under which 3712 you make payment to the third party based on the extent of your 3713 activity of conveying the work, and under which the third party 3714 grants, to any of the parties who would receive the covered work 3715 from you, a discriminatory patent license (a) in connection with 3716 copies of the covered work conveyed by you (or copies made from 3717 those copies), or (b) primarily for and in connection with specific 3718 products or compilations that contain the covered work, unless you 3719 entered into that arrangement, or that patent license was granted, 3720 prior to 28 March 2007. 3721 3722 Nothing in this License shall be construed as excluding or limiting 3723 any implied license or other defenses to infringement that may 3724 otherwise be available to you under applicable patent law. 3725 3726 12. No Surrender of Others' Freedom. 3727 3728 If conditions are imposed on you (whether by court order, agreement 3729 or otherwise) that contradict the conditions of this License, they 3730 do not excuse you from the conditions of this License. If you 3731 cannot convey a covered work so as to satisfy simultaneously your 3732 obligations under this License and any other pertinent obligations, 3733 then as a consequence you may not convey it at all. For example, 3734 if you agree to terms that obligate you to collect a royalty for 3735 further conveying from those to whom you convey the Program, the 3736 only way you could satisfy both those terms and this License would 3737 be to refrain entirely from conveying the Program. 3738 3739 13. Use with the GNU Affero General Public License. 3740 3741 Notwithstanding any other provision of this License, you have 3742 permission to link or combine any covered work with a work licensed 3743 under version 3 of the GNU Affero General Public License into a 3744 single combined work, and to convey the resulting work. The terms 3745 of this License will continue to apply to the part which is the 3746 covered work, but the special requirements of the GNU Affero 3747 General Public License, section 13, concerning interaction through 3748 a network will apply to the combination as such. 3749 3750 14. Revised Versions of this License. 3751 3752 The Free Software Foundation may publish revised and/or new 3753 versions of the GNU General Public License from time to time. Such 3754 new versions will be similar in spirit to the present version, but 3755 may differ in detail to address new problems or concerns. 3756 3757 Each version is given a distinguishing version number. If the 3758 Program specifies that a certain numbered version of the GNU 3759 General Public License "or any later version" applies to it, you 3760 have the option of following the terms and conditions either of 3761 that numbered version or of any later version published by the Free 3762 Software Foundation. If the Program does not specify a version 3763 number of the GNU General Public License, you may choose any 3764 version ever published by the Free Software Foundation. 3765 3766 If the Program specifies that a proxy can decide which future 3767 versions of the GNU General Public License can be used, that 3768 proxy's public statement of acceptance of a version permanently 3769 authorizes you to choose that version for the Program. 3770 3771 Later license versions may give you additional or different 3772 permissions. However, no additional obligations are imposed on any 3773 author or copyright holder as a result of your choosing to follow a 3774 later version. 3775 3776 15. Disclaimer of Warranty. 3777 3778 THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY 3779 APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE 3780 COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" 3781 WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, 3782 INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF 3783 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE 3784 RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. 3785 SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL 3786 NECESSARY SERVICING, REPAIR OR CORRECTION. 3787 3788 16. Limitation of Liability. 3789 3790 IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN 3791 WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES 3792 AND/OR CONVEYS THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR 3793 DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR 3794 CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE 3795 THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA 3796 BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD 3797 PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER 3798 PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF 3799 THE POSSIBILITY OF SUCH DAMAGES. 3800 3801 17. Interpretation of Sections 15 and 16. 3802 3803 If the disclaimer of warranty and limitation of liability provided 3804 above cannot be given local legal effect according to their terms, 3805 reviewing courts shall apply local law that most closely 3806 approximates an absolute waiver of all civil liability in 3807 connection with the Program, unless a warranty or assumption of 3808 liability accompanies a copy of the Program in return for a fee. 3809 3810 END OF TERMS AND CONDITIONS 3811 =========================== 3812 3813 How to Apply These Terms to Your New Programs 3814 ============================================= 3815 3816 If you develop a new program, and you want it to be of the greatest 3817 possible use to the public, the best way to achieve this is to make it 3818 free software which everyone can redistribute and change under these 3819 terms. 3820 3821 To do so, attach the following notices to the program. It is safest 3822 to attach them to the start of each source file to most effectively 3823 state the exclusion of warranty; and each file should have at least the 3824 "copyright" line and a pointer to where the full notice is found. 3825 3826 ONE LINE TO GIVE THE PROGRAM'S NAME AND A BRIEF IDEA OF WHAT IT DOES. 3827 Copyright (C) YEAR NAME OF AUTHOR 3828 3829 This program is free software: you can redistribute it and/or modify 3830 it under the terms of the GNU General Public License as published by 3831 the Free Software Foundation, either version 3 of the License, or (at 3832 your option) any later version. 3833 3834 This program is distributed in the hope that it will be useful, but 3835 WITHOUT ANY WARRANTY; without even the implied warranty of 3836 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 3837 General Public License for more details. 3838 3839 You should have received a copy of the GNU General Public License 3840 along with this program. If not, see <http://www.gnu.org/licenses/>. 3841 3842 Also add information on how to contact you by electronic and paper 3843 mail. 3844 3845 If the program does terminal interaction, make it output a short 3846 notice like this when it starts in an interactive mode: 3847 3848 PROGRAM Copyright (C) YEAR NAME OF AUTHOR 3849 This program comes with ABSOLUTELY NO WARRANTY; for details type 'show w'. 3850 This is free software, and you are welcome to redistribute it 3851 under certain conditions; type 'show c' for details. 3852 3853 The hypothetical commands 'show w' and 'show c' should show the 3854 appropriate parts of the General Public License. Of course, your 3855 program's commands might be different; for a GUI interface, you would 3856 use an "about box". 3857 3858 You should also get your employer (if you work as a programmer) or 3859 school, if any, to sign a "copyright disclaimer" for the program, if 3860 necessary. For more information on this, and how to apply and follow 3861 the GNU GPL, see <http://www.gnu.org/licenses/>. 3862 3863 The GNU General Public License does not permit incorporating your 3864 program into proprietary programs. If your program is a subroutine 3865 library, you may consider it more useful to permit linking proprietary 3866 applications with the library. If this is what you want to do, use the 3867 GNU Lesser General Public License instead of this License. But first, 3868 please read <http://www.gnu.org/philosophy/why-not-lgpl.html>. 3869 3870 3871 File: libgomp.info, Node: GNU Free Documentation License, Next: Funding, Prev: Copying, Up: Top 3872 3873 GNU Free Documentation License 3874 ****************************** 3875 3876 Version 1.3, 3 November 2008 3877 3878 Copyright (C) 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc. 3879 <http://fsf.org/> 3880 3881 Everyone is permitted to copy and distribute verbatim copies 3882 of this license document, but changing it is not allowed. 3883 3884 0. PREAMBLE 3885 3886 The purpose of this License is to make a manual, textbook, or other 3887 functional and useful document "free" in the sense of freedom: to 3888 assure everyone the effective freedom to copy and redistribute it, 3889 with or without modifying it, either commercially or 3890 noncommercially. Secondarily, this License preserves for the 3891 author and publisher a way to get credit for their work, while not 3892 being considered responsible for modifications made by others. 3893 3894 This License is a kind of "copyleft", which means that derivative 3895 works of the document must themselves be free in the same sense. 3896 It complements the GNU General Public License, which is a copyleft 3897 license designed for free software. 3898 3899 We have designed this License in order to use it for manuals for 3900 free software, because free software needs free documentation: a 3901 free program should come with manuals providing the same freedoms 3902 that the software does. But this License is not limited to 3903 software manuals; it can be used for any textual work, regardless 3904 of subject matter or whether it is published as a printed book. We 3905 recommend this License principally for works whose purpose is 3906 instruction or reference. 3907 3908 1. APPLICABILITY AND DEFINITIONS 3909 3910 This License applies to any manual or other work, in any medium, 3911 that contains a notice placed by the copyright holder saying it can 3912 be distributed under the terms of this License. Such a notice 3913 grants a world-wide, royalty-free license, unlimited in duration, 3914 to use that work under the conditions stated herein. The 3915 "Document", below, refers to any such manual or work. Any member 3916 of the public is a licensee, and is addressed as "you". You accept 3917 the license if you copy, modify or distribute the work in a way 3918 requiring permission under copyright law. 3919 3920 A "Modified Version" of the Document means any work containing the 3921 Document or a portion of it, either copied verbatim, or with 3922 modifications and/or translated into another language. 3923 3924 A "Secondary Section" is a named appendix or a front-matter section 3925 of the Document that deals exclusively with the relationship of the 3926 publishers or authors of the Document to the Document's overall 3927 subject (or to related matters) and contains nothing that could 3928 fall directly within that overall subject. (Thus, if the Document 3929 is in part a textbook of mathematics, a Secondary Section may not 3930 explain any mathematics.) The relationship could be a matter of 3931 historical connection with the subject or with related matters, or 3932 of legal, commercial, philosophical, ethical or political position 3933 regarding them. 3934 3935 The "Invariant Sections" are certain Secondary Sections whose 3936 titles are designated, as being those of Invariant Sections, in the 3937 notice that says that the Document is released under this License. 3938 If a section does not fit the above definition of Secondary then it 3939 is not allowed to be designated as Invariant. The Document may 3940 contain zero Invariant Sections. If the Document does not identify 3941 any Invariant Sections then there are none. 3942 3943 The "Cover Texts" are certain short passages of text that are 3944 listed, as Front-Cover Texts or Back-Cover Texts, in the notice 3945 that says that the Document is released under this License. A 3946 Front-Cover Text may be at most 5 words, and a Back-Cover Text may 3947 be at most 25 words. 3948 3949 A "Transparent" copy of the Document means a machine-readable copy, 3950 represented in a format whose specification is available to the 3951 general public, that is suitable for revising the document 3952 straightforwardly with generic text editors or (for images composed 3953 of pixels) generic paint programs or (for drawings) some widely 3954 available drawing editor, and that is suitable for input to text 3955 formatters or for automatic translation to a variety of formats 3956 suitable for input to text formatters. A copy made in an otherwise 3957 Transparent file format whose markup, or absence of markup, has 3958 been arranged to thwart or discourage subsequent modification by 3959 readers is not Transparent. An image format is not Transparent if 3960 used for any substantial amount of text. A copy that is not 3961 "Transparent" is called "Opaque". 3962 3963 Examples of suitable formats for Transparent copies include plain 3964 ASCII without markup, Texinfo input format, LaTeX input format, 3965 SGML or XML using a publicly available DTD, and standard-conforming 3966 simple HTML, PostScript or PDF designed for human modification. 3967 Examples of transparent image formats include PNG, XCF and JPG. 3968 Opaque formats include proprietary formats that can be read and 3969 edited only by proprietary word processors, SGML or XML for which 3970 the DTD and/or processing tools are not generally available, and 3971 the machine-generated HTML, PostScript or PDF produced by some word 3972 processors for output purposes only. 3973 3974 The "Title Page" means, for a printed book, the title page itself, 3975 plus such following pages as are needed to hold, legibly, the 3976 material this License requires to appear in the title page. For 3977 works in formats which do not have any title page as such, "Title 3978 Page" means the text near the most prominent appearance of the 3979 work's title, preceding the beginning of the body of the text. 3980 3981 The "publisher" means any person or entity that distributes copies 3982 of the Document to the public. 3983 3984 A section "Entitled XYZ" means a named subunit of the Document 3985 whose title either is precisely XYZ or contains XYZ in parentheses 3986 following text that translates XYZ in another language. (Here XYZ 3987 stands for a specific section name mentioned below, such as 3988 "Acknowledgements", "Dedications", "Endorsements", or "History".) 3989 To "Preserve the Title" of such a section when you modify the 3990 Document means that it remains a section "Entitled XYZ" according 3991 to this definition. 3992 3993 The Document may include Warranty Disclaimers next to the notice 3994 which states that this License applies to the Document. These 3995 Warranty Disclaimers are considered to be included by reference in 3996 this License, but only as regards disclaiming warranties: any other 3997 implication that these Warranty Disclaimers may have is void and 3998 has no effect on the meaning of this License. 3999 4000 2. VERBATIM COPYING 4001 4002 You may copy and distribute the Document in any medium, either 4003 commercially or noncommercially, provided that this License, the 4004 copyright notices, and the license notice saying this License 4005 applies to the Document are reproduced in all copies, and that you 4006 add no other conditions whatsoever to those of this License. You 4007 may not use technical measures to obstruct or control the reading 4008 or further copying of the copies you make or distribute. However, 4009 you may accept compensation in exchange for copies. If you 4010 distribute a large enough number of copies you must also follow the 4011 conditions in section 3. 4012 4013 You may also lend copies, under the same conditions stated above, 4014 and you may publicly display copies. 4015 4016 3. COPYING IN QUANTITY 4017 4018 If you publish printed copies (or copies in media that commonly 4019 have printed covers) of the Document, numbering more than 100, and 4020 the Document's license notice requires Cover Texts, you must 4021 enclose the copies in covers that carry, clearly and legibly, all 4022 these Cover Texts: Front-Cover Texts on the front cover, and 4023 Back-Cover Texts on the back cover. Both covers must also clearly 4024 and legibly identify you as the publisher of these copies. The 4025 front cover must present the full title with all words of the title 4026 equally prominent and visible. You may add other material on the 4027 covers in addition. Copying with changes limited to the covers, as 4028 long as they preserve the title of the Document and satisfy these 4029 conditions, can be treated as verbatim copying in other respects. 4030 4031 If the required texts for either cover are too voluminous to fit 4032 legibly, you should put the first ones listed (as many as fit 4033 reasonably) on the actual cover, and continue the rest onto 4034 adjacent pages. 4035 4036 If you publish or distribute Opaque copies of the Document 4037 numbering more than 100, you must either include a machine-readable 4038 Transparent copy along with each Opaque copy, or state in or with 4039 each Opaque copy a computer-network location from which the general 4040 network-using public has access to download using public-standard 4041 network protocols a complete Transparent copy of the Document, free 4042 of added material. If you use the latter option, you must take 4043 reasonably prudent steps, when you begin distribution of Opaque 4044 copies in quantity, to ensure that this Transparent copy will 4045 remain thus accessible at the stated location until at least one 4046 year after the last time you distribute an Opaque copy (directly or 4047 through your agents or retailers) of that edition to the public. 4048 4049 It is requested, but not required, that you contact the authors of 4050 the Document well before redistributing any large number of copies, 4051 to give them a chance to provide you with an updated version of the 4052 Document. 4053 4054 4. MODIFICATIONS 4055 4056 You may copy and distribute a Modified Version of the Document 4057 under the conditions of sections 2 and 3 above, provided that you 4058 release the Modified Version under precisely this License, with the 4059 Modified Version filling the role of the Document, thus licensing 4060 distribution and modification of the Modified Version to whoever 4061 possesses a copy of it. In addition, you must do these things in 4062 the Modified Version: 4063 4064 A. Use in the Title Page (and on the covers, if any) a title 4065 distinct from that of the Document, and from those of previous 4066 versions (which should, if there were any, be listed in the 4067 History section of the Document). You may use the same title 4068 as a previous version if the original publisher of that 4069 version gives permission. 4070 4071 B. List on the Title Page, as authors, one or more persons or 4072 entities responsible for authorship of the modifications in 4073 the Modified Version, together with at least five of the 4074 principal authors of the Document (all of its principal 4075 authors, if it has fewer than five), unless they release you 4076 from this requirement. 4077 4078 C. State on the Title page the name of the publisher of the 4079 Modified Version, as the publisher. 4080 4081 D. Preserve all the copyright notices of the Document. 4082 4083 E. Add an appropriate copyright notice for your modifications 4084 adjacent to the other copyright notices. 4085 4086 F. Include, immediately after the copyright notices, a license 4087 notice giving the public permission to use the Modified 4088 Version under the terms of this License, in the form shown in 4089 the Addendum below. 4090 4091 G. Preserve in that license notice the full lists of Invariant 4092 Sections and required Cover Texts given in the Document's 4093 license notice. 4094 4095 H. Include an unaltered copy of this License. 4096 4097 I. Preserve the section Entitled "History", Preserve its Title, 4098 and add to it an item stating at least the title, year, new 4099 authors, and publisher of the Modified Version as given on the 4100 Title Page. If there is no section Entitled "History" in the 4101 Document, create one stating the title, year, authors, and 4102 publisher of the Document as given on its Title Page, then add 4103 an item describing the Modified Version as stated in the 4104 previous sentence. 4105 4106 J. Preserve the network location, if any, given in the Document 4107 for public access to a Transparent copy of the Document, and 4108 likewise the network locations given in the Document for 4109 previous versions it was based on. These may be placed in the 4110 "History" section. You may omit a network location for a work 4111 that was published at least four years before the Document 4112 itself, or if the original publisher of the version it refers 4113 to gives permission. 4114 4115 K. For any section Entitled "Acknowledgements" or "Dedications", 4116 Preserve the Title of the section, and preserve in the section 4117 all the substance and tone of each of the contributor 4118 acknowledgements and/or dedications given therein. 4119 4120 L. Preserve all the Invariant Sections of the Document, unaltered 4121 in their text and in their titles. Section numbers or the 4122 equivalent are not considered part of the section titles. 4123 4124 M. Delete any section Entitled "Endorsements". Such a section 4125 may not be included in the Modified Version. 4126 4127 N. Do not retitle any existing section to be Entitled 4128 "Endorsements" or to conflict in title with any Invariant 4129 Section. 4130 4131 O. Preserve any Warranty Disclaimers. 4132 4133 If the Modified Version includes new front-matter sections or 4134 appendices that qualify as Secondary Sections and contain no 4135 material copied from the Document, you may at your option designate 4136 some or all of these sections as invariant. To do this, add their 4137 titles to the list of Invariant Sections in the Modified Version's 4138 license notice. These titles must be distinct from any other 4139 section titles. 4140 4141 You may add a section Entitled "Endorsements", provided it contains 4142 nothing but endorsements of your Modified Version by various 4143 parties--for example, statements of peer review or that the text 4144 has been approved by an organization as the authoritative 4145 definition of a standard. 4146 4147 You may add a passage of up to five words as a Front-Cover Text, 4148 and a passage of up to 25 words as a Back-Cover Text, to the end of 4149 the list of Cover Texts in the Modified Version. Only one passage 4150 of Front-Cover Text and one of Back-Cover Text may be added by (or 4151 through arrangements made by) any one entity. If the Document 4152 already includes a cover text for the same cover, previously added 4153 by you or by arrangement made by the same entity you are acting on 4154 behalf of, you may not add another; but you may replace the old 4155 one, on explicit permission from the previous publisher that added 4156 the old one. 4157 4158 The author(s) and publisher(s) of the Document do not by this 4159 License give permission to use their names for publicity for or to 4160 assert or imply endorsement of any Modified Version. 4161 4162 5. COMBINING DOCUMENTS 4163 4164 You may combine the Document with other documents released under 4165 this License, under the terms defined in section 4 above for 4166 modified versions, provided that you include in the combination all 4167 of the Invariant Sections of all of the original documents, 4168 unmodified, and list them all as Invariant Sections of your 4169 combined work in its license notice, and that you preserve all 4170 their Warranty Disclaimers. 4171 4172 The combined work need only contain one copy of this License, and 4173 multiple identical Invariant Sections may be replaced with a single 4174 copy. If there are multiple Invariant Sections with the same name 4175 but different contents, make the title of each such section unique 4176 by adding at the end of it, in parentheses, the name of the 4177 original author or publisher of that section if known, or else a 4178 unique number. Make the same adjustment to the section titles in 4179 the list of Invariant Sections in the license notice of the 4180 combined work. 4181 4182 In the combination, you must combine any sections Entitled 4183 "History" in the various original documents, forming one section 4184 Entitled "History"; likewise combine any sections Entitled 4185 "Acknowledgements", and any sections Entitled "Dedications". You 4186 must delete all sections Entitled "Endorsements." 4187 4188 6. COLLECTIONS OF DOCUMENTS 4189 4190 You may make a collection consisting of the Document and other 4191 documents released under this License, and replace the individual 4192 copies of this License in the various documents with a single copy 4193 that is included in the collection, provided that you follow the 4194 rules of this License for verbatim copying of each of the documents 4195 in all other respects. 4196 4197 You may extract a single document from such a collection, and 4198 distribute it individually under this License, provided you insert 4199 a copy of this License into the extracted document, and follow this 4200 License in all other respects regarding verbatim copying of that 4201 document. 4202 4203 7. AGGREGATION WITH INDEPENDENT WORKS 4204 4205 A compilation of the Document or its derivatives with other 4206 separate and independent documents or works, in or on a volume of a 4207 storage or distribution medium, is called an "aggregate" if the 4208 copyright resulting from the compilation is not used to limit the 4209 legal rights of the compilation's users beyond what the individual 4210 works permit. When the Document is included in an aggregate, this 4211 License does not apply to the other works in the aggregate which 4212 are not themselves derivative works of the Document. 4213 4214 If the Cover Text requirement of section 3 is applicable to these 4215 copies of the Document, then if the Document is less than one half 4216 of the entire aggregate, the Document's Cover Texts may be placed 4217 on covers that bracket the Document within the aggregate, or the 4218 electronic equivalent of covers if the Document is in electronic 4219 form. Otherwise they must appear on printed covers that bracket 4220 the whole aggregate. 4221 4222 8. TRANSLATION 4223 4224 Translation is considered a kind of modification, so you may 4225 distribute translations of the Document under the terms of section 4226 4. Replacing Invariant Sections with translations requires special 4227 permission from their copyright holders, but you may include 4228 translations of some or all Invariant Sections in addition to the 4229 original versions of these Invariant Sections. You may include a 4230 translation of this License, and all the license notices in the 4231 Document, and any Warranty Disclaimers, provided that you also 4232 include the original English version of this License and the 4233 original versions of those notices and disclaimers. In case of a 4234 disagreement between the translation and the original version of 4235 this License or a notice or disclaimer, the original version will 4236 prevail. 4237 4238 If a section in the Document is Entitled "Acknowledgements", 4239 "Dedications", or "History", the requirement (section 4) to 4240 Preserve its Title (section 1) will typically require changing the 4241 actual title. 4242 4243 9. TERMINATION 4244 4245 You may not copy, modify, sublicense, or distribute the Document 4246 except as expressly provided under this License. Any attempt 4247 otherwise to copy, modify, sublicense, or distribute it is void, 4248 and will automatically terminate your rights under this License. 4249 4250 However, if you cease all violation of this License, then your 4251 license from a particular copyright holder is reinstated (a) 4252 provisionally, unless and until the copyright holder explicitly and 4253 finally terminates your license, and (b) permanently, if the 4254 copyright holder fails to notify you of the violation by some 4255 reasonable means prior to 60 days after the cessation. 4256 4257 Moreover, your license from a particular copyright holder is 4258 reinstated permanently if the copyright holder notifies you of the 4259 violation by some reasonable means, this is the first time you have 4260 received notice of violation of this License (for any work) from 4261 that copyright holder, and you cure the violation prior to 30 days 4262 after your receipt of the notice. 4263 4264 Termination of your rights under this section does not terminate 4265 the licenses of parties who have received copies or rights from you 4266 under this License. If your rights have been terminated and not 4267 permanently reinstated, receipt of a copy of some or all of the 4268 same material does not give you any rights to use it. 4269 4270 10. FUTURE REVISIONS OF THIS LICENSE 4271 4272 The Free Software Foundation may publish new, revised versions of 4273 the GNU Free Documentation License from time to time. Such new 4274 versions will be similar in spirit to the present version, but may 4275 differ in detail to address new problems or concerns. See 4276 <http://www.gnu.org/copyleft/>. 4277 4278 Each version of the License is given a distinguishing version 4279 number. If the Document specifies that a particular numbered 4280 version of this License "or any later version" applies to it, you 4281 have the option of following the terms and conditions either of 4282 that specified version or of any later version that has been 4283 published (not as a draft) by the Free Software Foundation. If the 4284 Document does not specify a version number of this License, you may 4285 choose any version ever published (not as a draft) by the Free 4286 Software Foundation. If the Document specifies that a proxy can 4287 decide which future versions of this License can be used, that 4288 proxy's public statement of acceptance of a version permanently 4289 authorizes you to choose that version for the Document. 4290 4291 11. RELICENSING 4292 4293 "Massive Multiauthor Collaboration Site" (or "MMC Site") means any 4294 World Wide Web server that publishes copyrightable works and also 4295 provides prominent facilities for anybody to edit those works. A 4296 public wiki that anybody can edit is an example of such a server. 4297 A "Massive Multiauthor Collaboration" (or "MMC") contained in the 4298 site means any set of copyrightable works thus published on the MMC 4299 site. 4300 4301 "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0 4302 license published by Creative Commons Corporation, a not-for-profit 4303 corporation with a principal place of business in San Francisco, 4304 California, as well as future copyleft versions of that license 4305 published by that same organization. 4306 4307 "Incorporate" means to publish or republish a Document, in whole or 4308 in part, as part of another Document. 4309 4310 An MMC is "eligible for relicensing" if it is licensed under this 4311 License, and if all works that were first published under this 4312 License somewhere other than this MMC, and subsequently 4313 incorporated in whole or in part into the MMC, (1) had no cover 4314 texts or invariant sections, and (2) were thus incorporated prior 4315 to November 1, 2008. 4316 4317 The operator of an MMC Site may republish an MMC contained in the 4318 site under CC-BY-SA on the same site at any time before August 1, 4319 2009, provided the MMC is eligible for relicensing. 4320 4321 ADDENDUM: How to use this License for your documents 4322 ==================================================== 4323 4324 To use this License in a document you have written, include a copy of 4325 the License in the document and put the following copyright and license 4326 notices just after the title page: 4327 4328 Copyright (C) YEAR YOUR NAME. 4329 Permission is granted to copy, distribute and/or modify this document 4330 under the terms of the GNU Free Documentation License, Version 1.3 4331 or any later version published by the Free Software Foundation; 4332 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover 4333 Texts. A copy of the license is included in the section entitled ``GNU 4334 Free Documentation License''. 4335 4336 If you have Invariant Sections, Front-Cover Texts and Back-Cover 4337 Texts, replace the "with...Texts." line with this: 4338 4339 with the Invariant Sections being LIST THEIR TITLES, with 4340 the Front-Cover Texts being LIST, and with the Back-Cover Texts 4341 being LIST. 4342 4343 If you have Invariant Sections without Cover Texts, or some other 4344 combination of the three, merge those two alternatives to suit the 4345 situation. 4346 4347 If your document contains nontrivial examples of program code, we 4348 recommend releasing these examples in parallel under your choice of free 4349 software license, such as the GNU General Public License, to permit 4350 their use in free software. 4351 4352 4353 File: libgomp.info, Node: Funding, Next: Library Index, Prev: GNU Free Documentation License, Up: Top 4354 4355 Funding Free Software 4356 ********************* 4357 4358 If you want to have more free software a few years from now, it makes 4359 sense for you to help encourage people to contribute funds for its 4360 development. The most effective approach known is to encourage 4361 commercial redistributors to donate. 4362 4363 Users of free software systems can boost the pace of development by 4364 encouraging for-a-fee distributors to donate part of their selling price 4365 to free software developers--the Free Software Foundation, and others. 4366 4367 The way to convince distributors to do this is to demand it and 4368 expect it from them. So when you compare distributors, judge them 4369 partly by how much they give to free software development. Show 4370 distributors they must compete to be the one who gives the most. 4371 4372 To make this approach work, you must insist on numbers that you can 4373 compare, such as, "We will donate ten dollars to the Frobnitz project 4374 for each disk sold." Don't be satisfied with a vague promise, such as 4375 "A portion of the profits are donated," since it doesn't give a basis 4376 for comparison. 4377 4378 Even a precise fraction "of the profits from this disk" is not very 4379 meaningful, since creative accounting and unrelated business decisions 4380 can greatly alter what fraction of the sales price counts as profit. If 4381 the price you pay is $50, ten percent of the profit is probably less 4382 than a dollar; it might be a few cents, or nothing at all. 4383 4384 Some redistributors do development work themselves. This is useful 4385 too; but to keep everyone honest, you need to inquire how much they do, 4386 and what kind. Some kinds of development make much more long-term 4387 difference than others. For example, maintaining a separate version of 4388 a program contributes very little; maintaining the standard version of a 4389 program for the whole community contributes much. Easy new ports 4390 contribute little, since someone else would surely do them; difficult 4391 ports such as adding a new CPU to the GNU Compiler Collection contribute 4392 more; major new features or packages contribute the most. 4393 4394 By establishing the idea that supporting further development is "the 4395 proper thing to do" when distributing free software for a fee, we can 4396 assure a steady flow of resources into making more free software. 4397 4398 Copyright (C) 1994 Free Software Foundation, Inc. 4399 Verbatim copying and redistribution of this section is permitted 4400 without royalty; alteration is not permitted. 4401 4402 4403 File: libgomp.info, Node: Library Index, Prev: Funding, Up: Top 4404 4405 Library Index 4406 ************* 4407 4408 [index] 4409 * Menu: 4410 4411 * Environment Variable: OMP_CANCELLATION. (line 6) 4412 * Environment Variable <1>: OMP_DISPLAY_ENV. (line 6) 4413 * Environment Variable <2>: OMP_DEFAULT_DEVICE. (line 6) 4414 * Environment Variable <3>: OMP_DYNAMIC. (line 6) 4415 * Environment Variable <4>: OMP_MAX_ACTIVE_LEVELS. (line 6) 4416 * Environment Variable <5>: OMP_MAX_TASK_PRIORITY. (line 6) 4417 * Environment Variable <6>: OMP_NESTED. (line 6) 4418 * Environment Variable <7>: OMP_NUM_THREADS. (line 6) 4419 * Environment Variable <8>: OMP_PROC_BIND. (line 6) 4420 * Environment Variable <9>: OMP_PLACES. (line 6) 4421 * Environment Variable <10>: OMP_STACKSIZE. (line 6) 4422 * Environment Variable <11>: OMP_SCHEDULE. (line 6) 4423 * Environment Variable <12>: OMP_THREAD_LIMIT. (line 6) 4424 * Environment Variable <13>: OMP_WAIT_POLICY. (line 6) 4425 * Environment Variable <14>: GOMP_CPU_AFFINITY. (line 6) 4426 * Environment Variable <15>: GOMP_DEBUG. (line 6) 4427 * Environment Variable <16>: GOMP_STACKSIZE. (line 6) 4428 * Environment Variable <17>: GOMP_SPINCOUNT. (line 6) 4429 * Environment Variable <18>: GOMP_RTEMS_THREAD_POOLS. 4430 (line 6) 4431 * FDL, GNU Free Documentation License: GNU Free Documentation License. 4432 (line 6) 4433 * Implementation specific setting: OMP_NESTED. (line 6) 4434 * Implementation specific setting <1>: OMP_NUM_THREADS. (line 6) 4435 * Implementation specific setting <2>: OMP_SCHEDULE. (line 6) 4436 * Implementation specific setting <3>: GOMP_STACKSIZE. (line 6) 4437 * Implementation specific setting <4>: GOMP_SPINCOUNT. (line 6) 4438 * Implementation specific setting <5>: GOMP_RTEMS_THREAD_POOLS. 4439 (line 6) 4440 * Introduction: Top. (line 6) 4441 4442 4443 4444 Tag Table: 4445 Node: Top2083 4446 Node: Enabling OpenMP4613 4447 Node: Runtime Library Routines5401 4448 Node: omp_get_active_level8478 4449 Node: omp_get_ancestor_thread_num9178 4450 Node: omp_get_cancellation10108 4451 Node: omp_get_default_device10922 4452 Node: omp_get_dynamic11598 4453 Node: omp_get_level12473 4454 Node: omp_get_max_active_levels13093 4455 Node: omp_get_max_task_priority13798 4456 Node: omp_get_max_threads14418 4457 Node: omp_get_nested15175 4458 Node: omp_get_num_devices16089 4459 Node: omp_get_num_procs16610 4460 Node: omp_get_num_teams17149 4461 Node: omp_get_num_threads17665 4462 Node: omp_get_proc_bind18754 4463 Node: omp_get_schedule19675 4464 Node: omp_get_team_num20629 4465 Node: omp_get_team_size21128 4466 Node: omp_get_thread_limit22088 4467 Node: omp_get_thread_num22707 4468 Node: omp_in_parallel23578 4469 Node: omp_in_final24227 4470 Node: omp_is_initial_device24901 4471 Node: omp_set_default_device25594 4472 Node: omp_set_dynamic26385 4473 Node: omp_set_max_active_levels27271 4474 Node: omp_set_nested28048 4475 Node: omp_set_num_threads28940 4476 Node: omp_set_schedule29808 4477 Node: omp_init_lock30889 4478 Node: omp_set_lock31542 4479 Node: omp_test_lock32397 4480 Node: omp_unset_lock33373 4481 Node: omp_destroy_lock34304 4482 Node: omp_init_nest_lock34981 4483 Node: omp_set_nest_lock35716 4484 Node: omp_test_nest_lock36631 4485 Node: omp_unset_nest_lock37658 4486 Node: omp_destroy_nest_lock38673 4487 Node: omp_get_wtick39424 4488 Node: omp_get_wtime40016 4489 Node: Environment Variables40792 4490 Node: OMP_CANCELLATION42361 4491 Node: OMP_DISPLAY_ENV42894 4492 Node: OMP_DEFAULT_DEVICE43597 4493 Node: OMP_DYNAMIC44377 4494 Node: OMP_MAX_ACTIVE_LEVELS44973 4495 Node: OMP_MAX_TASK_PRIORITY45623 4496 Node: OMP_NESTED46281 4497 Node: OMP_NUM_THREADS46886 4498 Node: OMP_PROC_BIND47574 4499 Node: OMP_PLACES48766 4500 Node: OMP_STACKSIZE50943 4501 Node: OMP_SCHEDULE51767 4502 Node: OMP_THREAD_LIMIT52465 4503 Node: OMP_WAIT_POLICY53065 4504 Node: GOMP_CPU_AFFINITY53757 4505 Node: GOMP_DEBUG55487 4506 Node: GOMP_STACKSIZE55994 4507 Node: GOMP_SPINCOUNT56823 4508 Node: GOMP_RTEMS_THREAD_POOLS58027 4509 Node: Enabling OpenACC60205 4510 Node: OpenACC Runtime Library Routines61198 4511 Node: acc_get_num_devices64992 4512 Node: acc_set_device_type65718 4513 Node: acc_get_device_type66482 4514 Node: acc_set_device_num67197 4515 Node: acc_get_device_num67981 4516 Node: acc_async_test68778 4517 Node: acc_async_test_all69770 4518 Node: acc_wait70670 4519 Node: acc_wait_all71532 4520 Node: acc_wait_all_async72291 4521 Node: acc_wait_async73043 4522 Node: acc_init73750 4523 Node: acc_shutdown74396 4524 Node: acc_on_device75064 4525 Node: acc_malloc76068 4526 Node: acc_free76567 4527 Node: acc_copyin76994 4528 Node: acc_present_or_copyin78104 4529 Node: acc_create79730 4530 Node: acc_present_or_create80885 4531 Node: acc_copyout82519 4532 Node: acc_delete83542 4533 Node: acc_update_device84517 4534 Node: acc_update_self85629 4535 Node: acc_map_data86733 4536 Node: acc_unmap_data87418 4537 Node: acc_deviceptr87939 4538 Node: acc_hostptr88509 4539 Node: acc_is_present89073 4540 Node: acc_memcpy_to_device90600 4541 Node: acc_memcpy_from_device91263 4542 Node: acc_get_current_cuda_device91947 4543 Node: acc_get_current_cuda_context92544 4544 Node: acc_get_cuda_stream93144 4545 Node: acc_set_cuda_stream93735 4546 Node: OpenACC Environment Variables94380 4547 Node: ACC_DEVICE_TYPE94839 4548 Node: ACC_DEVICE_NUM95075 4549 Node: GCC_ACC_NOTIFY95332 4550 Node: CUDA Streams Usage95555 4551 Ref: CUDA Streams Usage-Footnote-197455 4552 Node: OpenACC Library Interoperability97564 4553 Ref: OpenACC Library Interoperability-Footnote-1103920 4554 Ref: OpenACC Library Interoperability-Footnote-2104172 4555 Node: The libgomp ABI104380 4556 Node: Implementing MASTER construct105236 4557 Node: Implementing CRITICAL construct105650 4558 Node: Implementing ATOMIC construct106389 4559 Node: Implementing FLUSH construct106870 4560 Node: Implementing BARRIER construct107141 4561 Node: Implementing THREADPRIVATE construct107410 4562 Node: Implementing PRIVATE clause108063 4563 Node: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses108644 4564 Node: Implementing REDUCTION clause109968 4565 Node: Implementing PARALLEL construct110525 4566 Node: Implementing FOR construct111782 4567 Node: Implementing ORDERED construct113780 4568 Node: Implementing SECTIONS construct114086 4569 Node: Implementing SINGLE construct114852 4570 Node: Implementing OpenACC's PARALLEL construct115564 4571 Node: Reporting Bugs115822 4572 Node: Copying116184 4573 Node: GNU Free Documentation License153731 4574 Node: Funding178854 4575 Node: Library Index181380 4576 4577 End Tag Table 4578
Indexes created Sun Mar 22 00:23:16 UTC 2026