OpenCores
URL https://opencores.org/ocsvn/openrisc_me/openrisc_me/trunk

Subversion Repositories openrisc_me

[/] [openrisc/] [trunk/] [gnu-src/] [gcc-4.5.1/] [libgomp/] [libgomp.info] - Blame information for rev 309

Go to most recent revision | Details | Compare with Previous | View Log

Line No. Rev Author Line
1 273 jeremybenn
This is libgomp.info, produced by makeinfo version 4.12 from
2
/space/rguenther/gcc-4.5.1/gcc-4.5.1/libgomp/libgomp.texi.
3
 
4
Copyright (C) 2006, 2007, 2008 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 1.2 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
22
funds for GNU development.
23
 
24
INFO-DIR-SECTION GNU Libraries
25
START-INFO-DIR-ENTRY
26
* libgomp: (libgomp).                    GNU OpenMP runtime library
27
END-INFO-DIR-ENTRY
28
 
29
   This manual documents the GNU implementation of the OpenMP API for
30
multi-platform shared-memory parallel programming in C/C++ and Fortran.
31
 
32
   Published by the Free Software Foundation 51 Franklin Street, Fifth
33
Floor Boston, MA 02110-1301 USA
34
 
35
   Copyright (C) 2006, 2007, 2008 Free Software Foundation, Inc.
36
 
37
   Permission is granted to copy, distribute and/or modify this document
38
under the terms of the GNU Free Documentation License, Version 1.2 or
39
any later version published by the Free Software Foundation; with the
40
Invariant Sections being "Funding Free Software", the Front-Cover texts
41
being (a) (see below), and with the Back-Cover Texts being (b) (see
42
below).  A copy of the license is included in the section entitled "GNU
43
Free Documentation License".
44
 
45
   (a) The FSF's Front-Cover Text is:
46
 
47
   A GNU Manual
48
 
49
   (b) The FSF's Back-Cover Text is:
50
 
51
   You have freedom to copy and modify this GNU Manual, like GNU
52
software.  Copies published by the Free Software Foundation raise
53
funds for GNU development.
54
 
55

56
File: libgomp.info,  Node: Top,  Next: Enabling OpenMP,  Up: (dir)
57
 
58
Introduction
59
************
60
 
61
This manual documents the usage of libgomp, the GNU implementation of
62
the OpenMP (http://www.openmp.org) Application Programming Interface
63
(API) for multi-platform shared-memory parallel programming in C/C++
64
and Fortran.
65
 
66
* Menu:
67
 
68
* Enabling OpenMP::            How to enable OpenMP for your applications.
69
* Runtime Library Routines::   The OpenMP runtime application programming
70
                               interface.
71
* Environment Variables::      Influencing runtime behavior with environment
72
                               variables.
73
* The libgomp ABI::            Notes on the external ABI presented by libgomp.
74
* Reporting Bugs::             How to report bugs in GNU OpenMP.
75
* Copying::                    GNU general public license says
76
                               how you can copy and share libgomp.
77
* GNU Free Documentation License::
78
                               How you can copy and share this manual.
79
* Funding::                    How to help assure continued work for free
80
                               software.
81
* Index::                      Index of this documentation.
82
 
83

84
File: libgomp.info,  Node: Enabling OpenMP,  Next: Runtime Library Routines,  Prev: Top,  Up: Top
85
 
86
1 Enabling OpenMP
87
*****************
88
 
89
To activate the OpenMP extensions for C/C++ and Fortran, the
90
compile-time flag `-fopenmp' must be specified. This enables the OpenMP
91
directive `#pragma omp' in C/C++ and `!$omp' directives in free form,
92
`c$omp', `*$omp' and `!$omp' directives in fixed form, `!$' conditional
93
compilation sentinels in free form and `c$', `*$' and `!$' sentinels in
94
fixed form, for Fortran. The flag also arranges for automatic linking
95
of the OpenMP runtime library (*note Runtime Library Routines::).
96
 
97
   A complete description of all OpenMP directives accepted may be
98
found in the OpenMP Application Program Interface
99
(http://www.openmp.org) manual, version 3.0.
100
 
101

102
File: libgomp.info,  Node: Runtime Library Routines,  Next: Environment Variables,  Prev: Enabling OpenMP,  Up: Top
103
 
104
2 Runtime Library Routines
105
**************************
106
 
107
The runtime routines described here are defined by section 3 of the
108
OpenMP specifications in version 3.0. The routines are structured in
109
following three parts:
110
 
111
   Control threads, processors and the parallel environment.
112
 
113
* Menu:
114
 
115
* omp_get_active_level::        Number of active parallel regions
116
* omp_get_ancestor_thread_num:: Ancestor thread ID
117
* omp_get_dynamic::             Dynamic teams setting
118
* omp_get_level::               Number of parallel regions
119
* omp_get_max_active_levels::   Maximal number of active regions
120
* omp_get_max_threads::         Maximal number of threads of parallel region
121
* omp_get_nested::              Nested parallel regions
122
* omp_get_num_procs::           Number of processors online
123
* omp_get_num_threads::         Size of the active team
124
* omp_get_schedule::            Obtain the runtime scheduling method
125
* omp_get_team_size::           Number of threads in a team
126
* omp_get_thread_limit::        Maximal number of threads
127
* omp_get_thread_num::          Current thread ID
128
* omp_in_parallel::             Whether a parallel region is active
129
* omp_set_dynamic::             Enable/disable dynamic teams
130
* omp_set_max_active_levels::   Limits the number of active parallel regions
131
* omp_set_nested::              Enable/disable nested parallel regions
132
* omp_set_num_threads::         Set upper team size limit
133
* omp_set_schedule::            Set the runtime scheduling method
134
 
135
   Initialize, set, test, unset and destroy simple and nested locks.
136
 
137
* Menu:
138
 
139
* omp_init_lock::            Initialize simple lock
140
* omp_set_lock::             Wait for and set simple lock
141
* omp_test_lock::            Test and set simple lock if available
142
* omp_unset_lock::           Unset simple lock
143
* omp_destroy_lock::         Destroy simple lock
144
* omp_init_nest_lock::       Initialize nested lock
145
* omp_set_nest_lock::        Wait for and set simple lock
146
* omp_test_nest_lock::       Test and set nested lock if available
147
* omp_unset_nest_lock::      Unset nested lock
148
* omp_destroy_nest_lock::    Destroy nested lock
149
 
150
   Portable, thread-based, wall clock timer.
151
 
152
* Menu:
153
 
154
* omp_get_wtick::            Get timer precision.
155
* omp_get_wtime::            Elapsed wall clock time.
156
 
157

158
File: libgomp.info,  Node: omp_get_active_level,  Next: omp_get_ancestor_thread_num,  Up: Runtime Library Routines
159
 
160
2.1 `omp_get_active_level' - Number of parallel regions
161
=======================================================
162
 
163
_Description_:
164
     This function returns the nesting level for the active parallel
165
     blocks, which enclose the calling call.
166
 
167
_C/C++_
168
     _Prototype_:  `int omp_get_active_level();'
169
 
170
_Fortran_:
171
     _Interface_:  `integer omp_get_active_level()'
172
 
173
_See also_:
174
     *note omp_get_level::, *note omp_get_max_active_levels::, *note
175
     omp_set_max_active_levels::
176
 
177
_Reference_:
178
     OpenMP specifications v3.0 (http://www.openmp.org/), section
179
     3.2.19.
180
 
181

182
File: libgomp.info,  Node: omp_get_ancestor_thread_num,  Next: omp_get_dynamic,  Prev: omp_get_active_level,  Up: Runtime Library Routines
183
 
184
2.2 `omp_get_ancestor_thread_num' - Ancestor thread ID
185
======================================================
186
 
187
_Description_:
188
     This function returns the thread identification number for the
189
     given nesting level of the current thread. For values of LEVEL
190
     outside zero to `omp_get_level' -1 is returned; if LEVEL is
191
     `omp_get_level' the result is identical to `omp_get_thread_num'.
192
 
193
_C/C++_
194
     _Prototype_:  `int omp_get_ancestor_thread_num(int level);'
195
 
196
_Fortran_:
197
     _Interface_:  `integer omp_ancestor_thread_num(level)'
198
                   `integer level'
199
 
200
_See also_:
201
     *note omp_get_level::, *note omp_get_thread_num::, *note
202
     omp_get_team_size::
203
 
204
_Reference_:
205
     OpenMP specifications v3.0 (http://www.openmp.org/), section
206
     3.2.17.
207
 
208

209
File: libgomp.info,  Node: omp_get_dynamic,  Next: omp_get_level,  Prev: omp_get_ancestor_thread_num,  Up: Runtime Library Routines
210
 
211
2.3 `omp_get_dynamic' - Dynamic teams setting
212
=============================================
213
 
214
_Description_:
215
     This function returns `true' if enabled, `false' otherwise.  Here,
216
     `true' and `false' represent their language-specific counterparts.
217
 
218
     The dynamic team setting may be initialized at startup by the
219
     `OMP_DYNAMIC' environment variable or at runtime using
220
     `omp_set_dynamic'. If undefined, dynamic adjustment is disabled by
221
     default.
222
 
223
_C/C++_:
224
     _Prototype_:  `int omp_get_dynamic();'
225
 
226
_Fortran_:
227
     _Interface_:  `logical function omp_get_dynamic()'
228
 
229
_See also_:
230
     *note omp_set_dynamic::, *note OMP_DYNAMIC::
231
 
232
_Reference_:
233
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.8.
234
 
235

236
File: libgomp.info,  Node: omp_get_level,  Next: omp_get_max_active_levels,  Prev: omp_get_dynamic,  Up: Runtime Library Routines
237
 
238
2.4 `omp_get_level' - Obtain the current nesting level
239
======================================================
240
 
241
_Description_:
242
     This function returns the nesting level for the parallel blocks,
243
     which enclose the calling call.
244
 
245
_C/C++_
246
     _Prototype_:  `int omp_get level();'
247
 
248
_Fortran_:
249
     _Interface_:  `integer omp_level()'
250
 
251
_See also_:
252
     *note omp_get_active_level::
253
 
254
_Reference_:
255
     OpenMP specifications v3.0 (http://www.openmp.org/), section
256
     3.2.16.
257
 
258

259
File: libgomp.info,  Node: omp_get_max_active_levels,  Next: omp_get_max_threads,  Prev: omp_get_level,  Up: Runtime Library Routines
260
 
261
2.5 `omp_set_max_active_levels' - Maximal number of active regions
262
==================================================================
263
 
264
_Description_:
265
     This function obtains the maximally allowed number of nested,
266
     active parallel regions.
267
 
268
_C/C++_
269
     _Prototype_:  `int omp_get_max_active_levels();'
270
 
271
_Fortran_:
272
     _Interface_:  `int omp_get_max_active_levels()'
273
 
274
_See also_:
275
     *note omp_set_max_active_levels::, *note omp_get_active_level::
276
 
277
_Reference_:
278
     OpenMP specifications v3.0 (http://www.openmp.org/), section
279
     3.2.14.
280
 
281

282
File: libgomp.info,  Node: omp_get_max_threads,  Next: omp_get_nested,  Prev: omp_get_max_active_levels,  Up: Runtime Library Routines
283
 
284
2.6 `omp_get_max_threads' - Maximal number of threads of parallel region
285
========================================================================
286
 
287
_Description_:
288
     Return the maximal number of threads used for the current parallel
289
     region that does not use the clause `num_threads'.
290
 
291
_C/C++_:
292
     _Prototype_:  `int omp_get_max_threads();'
293
 
294
_Fortran_:
295
     _Interface_:  `integer function omp_get_max_threads()'
296
 
297
_See also_:
298
     *note omp_set_num_threads::, *note omp_set_dynamic::, *note
299
     omp_get_thread_limit::
300
 
301
_Reference_:
302
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.3.
303
 
304

305
File: libgomp.info,  Node: omp_get_nested,  Next: omp_get_num_procs,  Prev: omp_get_max_threads,  Up: Runtime Library Routines
306
 
307
2.7 `omp_get_nested' - Nested parallel regions
308
==============================================
309
 
310
_Description_:
311
     This function returns `true' if nested parallel regions are
312
     enabled, `false' otherwise. Here, `true' and `false' represent
313
     their language-specific counterparts.
314
 
315
     Nested parallel regions may be initialized at startup by the
316
     `OMP_NESTED' environment variable or at runtime using
317
     `omp_set_nested'. If undefined, nested parallel regions are
318
     disabled by default.
319
 
320
_C/C++_:
321
     _Prototype_:  `int omp_get_nested();'
322
 
323
_Fortran_:
324
     _Interface_:  `integer function omp_get_nested()'
325
 
326
_See also_:
327
     *note omp_set_nested::, *note OMP_NESTED::
328
 
329
_Reference_:
330
     OpenMP specifications v3.0 (http://www.openmp.org/), section
331
     3.2.10.
332
 
333

334
File: libgomp.info,  Node: omp_get_num_procs,  Next: omp_get_num_threads,  Prev: omp_get_nested,  Up: Runtime Library Routines
335
 
336
2.8 `omp_get_num_procs' - Number of processors online
337
=====================================================
338
 
339
_Description_:
340
     Returns the number of processors online.
341
 
342
_C/C++_:
343
     _Prototype_:  `int omp_get_num_procs();'
344
 
345
_Fortran_:
346
     _Interface_:  `integer function omp_get_num_procs()'
347
 
348
_Reference_:
349
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.5.
350
 
351

352
File: libgomp.info,  Node: omp_get_num_threads,  Next: omp_get_schedule,  Prev: omp_get_num_procs,  Up: Runtime Library Routines
353
 
354
2.9 `omp_get_num_threads' - Size of the active team
355
===================================================
356
 
357
_Description_:
358
     The number of threads in the current team. In a sequential section
359
     of the program `omp_get_num_threads' returns 1.
360
 
361
     The default team size may be initialized at startup by the
362
     `OMP_NUM_THREADS' environment variable. At runtime, the size of
363
     the current team may be set either by the `NUM_THREADS' clause or
364
     by `omp_set_num_threads'. If none of the above were used to define
365
     a specific value and `OMP_DYNAMIC' is disabled, one thread per CPU
366
     online is used.
367
 
368
_C/C++_:
369
     _Prototype_:  `int omp_get_num_threads();'
370
 
371
_Fortran_:
372
     _Interface_:  `integer function omp_get_num_threads()'
373
 
374
_See also_:
375
     *note omp_get_max_threads::, *note omp_set_num_threads::, *note
376
     OMP_NUM_THREADS::
377
 
378
_Reference_:
379
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.2.
380
 
381

382
File: libgomp.info,  Node: omp_get_schedule,  Next: omp_get_team_size,  Prev: omp_get_num_threads,  Up: Runtime Library Routines
383
 
384
2.10 `omp_get_schedule' - Obtain the runtime scheduling method
385
==============================================================
386
 
387
_Description_:
388
     Obtain runtime the scheduling method. The KIND argument will be
389
     set to the value `omp_sched_static', `omp_sched_dynamic',
390
     `opm_sched_guided' or `auto'. The second argument, MODIFIER, is
391
     set to the chunk size.
392
 
393
_C/C++_
394
     _Prototype_:  `omp_schedule(omp_sched_t * kind, int *modifier);'
395
 
396
_Fortran_:
397
     _Interface_:  `subroutine omp_schedule(kind, modifier)'
398
                   `integer(kind=omp_sched_kind) kind'
399
                   `integer modifier'
400
 
401
_See also_:
402
     *note omp_set_schedule::, *note OMP_SCHEDULE::
403
 
404
_Reference_:
405
     OpenMP specifications v3.0 (http://www.openmp.org/), section
406
     3.2.12.
407
 
408

409
File: libgomp.info,  Node: omp_get_team_size,  Next: omp_get_thread_limit,  Prev: omp_get_schedule,  Up: Runtime Library Routines
410
 
411
2.11 `omp_get_team_size' - Number of threads in a team
412
======================================================
413
 
414
_Description_:
415
     This function returns the number of threads in a thread team to
416
     which either the current thread or its ancestor belongs. For
417
     values of LEVEL outside zero to `omp_get_level' -1 is returned; if
418
     LEVEL is zero 1 is returned and for `omp_get_level' the result is
419
     identical to `omp_get_num_threads'.
420
 
421
_C/C++_:
422
     _Prototype_:  `int omp_get_time_size(int level);'
423
 
424
_Fortran_:
425
     _Interface_:  `integer function omp_get_team_size(level)'
426
                   `integer level'
427
 
428
_See also_:
429
     *note omp_get_num_threads::, *note omp_get_level::, *note
430
     omp_get_ancestor_thread_num::
431
 
432
_Reference_:
433
     OpenMP specifications v3.0 (http://www.openmp.org/), section
434
     3.2.18.
435
 
436

437
File: libgomp.info,  Node: omp_get_thread_limit,  Next: omp_get_thread_num,  Prev: omp_get_team_size,  Up: Runtime Library Routines
438
 
439
2.12 `omp_get_thread_limit' - Maximal number of threads
440
=======================================================
441
 
442
_Description_:
443
     Return the maximal number of threads of the program.
444
 
445
_C/C++_:
446
     _Prototype_:  `int omp_get_thread_limit();'
447
 
448
_Fortran_:
449
     _Interface_:  `integer function omp_get_thread_limit()'
450
 
451
_See also_:
452
     *note omp_get_max_threads::, *note OMP_THREAD_LIMIT::
453
 
454
_Reference_:
455
     OpenMP specifications v3.0 (http://www.openmp.org/), section
456
     3.2.13.
457
 
458

459
File: libgomp.info,  Node: omp_get_thread_num,  Next: omp_in_parallel,  Prev: omp_get_thread_limit,  Up: Runtime Library Routines
460
 
461
2.13 `omp_get_thread_num' - Current thread ID
462
=============================================
463
 
464
_Description_:
465
     Unique thread identification number within the current team.  In a
466
     sequential parts of the program, `omp_get_thread_num' always
467
     returns 0. In parallel regions the return value varies from 0 to
468
     `omp_get_num_threads'-1 inclusive. The return value of the master
469
     thread of a team is always 0.
470
 
471
_C/C++_:
472
     _Prototype_:  `int omp_get_thread_num();'
473
 
474
_Fortran_:
475
     _Interface_:  `integer function omp_get_thread_num()'
476
 
477
_See also_:
478
     *note omp_get_num_threads::, *note omp_get_ancestor_thread_num::
479
 
480
_Reference_:
481
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.4.
482
 
483

484
File: libgomp.info,  Node: omp_in_parallel,  Next: omp_set_dynamic,  Prev: omp_get_thread_num,  Up: Runtime Library Routines
485
 
486
2.14 `omp_in_parallel' - Whether a parallel region is active
487
============================================================
488
 
489
_Description_:
490
     This function returns `true' if currently running in parallel,
491
     `false' otherwise. Here, `true' and `false' represent their
492
     language-specific counterparts.
493
 
494
_C/C++_:
495
     _Prototype_:  `int omp_in_parallel();'
496
 
497
_Fortran_:
498
     _Interface_:  `logical function omp_in_parallel()'
499
 
500
_Reference_:
501
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.6.
502
 
503

504
File: libgomp.info,  Node: omp_set_dynamic,  Next: omp_set_max_active_levels,  Prev: omp_in_parallel,  Up: Runtime Library Routines
505
 
506
2.15 `omp_set_dynamic' - Enable/disable dynamic teams
507
=====================================================
508
 
509
_Description_:
510
     Enable or disable the dynamic adjustment of the number of threads
511
     within a team. The function takes the language-specific equivalent
512
     of `true' and `false', where `true' enables dynamic adjustment of
513
     team sizes and `false' disables it.
514
 
515
_C/C++_:
516
     _Prototype_:  `void omp_set_dynamic(int);'
517
 
518
_Fortran_:
519
     _Interface_:  `subroutine omp_set_dynamic(set)'
520
                   `integer, intent(in) :: set'
521
 
522
_See also_:
523
     *note OMP_DYNAMIC::, *note omp_get_dynamic::
524
 
525
_Reference_:
526
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.7.
527
 
528

529
File: libgomp.info,  Node: omp_set_max_active_levels,  Next: omp_set_nested,  Prev: omp_set_dynamic,  Up: Runtime Library Routines
530
 
531
2.16 `omp_set_max_active_levels' - Limits the number of active parallel regions
532
===============================================================================
533
 
534
_Description_:
535
     This function limits the maximally allowed number of nested,
536
     active parallel regions.
537
 
538
_C/C++_
539
     _Prototype_:  `omp_set_max_active_levels(int max_levels);'
540
 
541
_Fortran_:
542
     _Interface_:  `omp_max_active_levels(max_levels)'
543
                   `integer max_levels'
544
 
545
_See also_:
546
     *note omp_get_max_active_levels::, *note omp_get_active_level::
547
 
548
_Reference_:
549
     OpenMP specifications v3.0 (http://www.openmp.org/), section
550
     3.2.14.
551
 
552

553
File: libgomp.info,  Node: omp_set_nested,  Next: omp_set_num_threads,  Prev: omp_set_max_active_levels,  Up: Runtime Library Routines
554
 
555
2.17 `omp_set_nested' - Enable/disable nested parallel regions
556
==============================================================
557
 
558
_Description_:
559
     Enable or disable nested parallel regions, i.e., whether team
560
     members are allowed to create new teams. The function takes the
561
     language-specific equivalent of `true' and `false', where `true'
562
     enables dynamic adjustment of team sizes and `false' disables it.
563
 
564
_C/C++_:
565
     _Prototype_:  `void omp_set_dynamic(int);'
566
 
567
_Fortran_:
568
     _Interface_:  `subroutine omp_set_dynamic(set)'
569
                   `integer, intent(in) :: set'
570
 
571
_See also_:
572
     *note OMP_NESTED::, *note omp_get_nested::
573
 
574
_Reference_:
575
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.9.
576
 
577

578
File: libgomp.info,  Node: omp_set_num_threads,  Next: omp_set_schedule,  Prev: omp_set_nested,  Up: Runtime Library Routines
579
 
580
2.18 `omp_set_num_threads' - Set upper team size limit
581
======================================================
582
 
583
_Description_:
584
     Specifies the number of threads used by default in subsequent
585
     parallel sections, if those do not specify a `num_threads' clause.
586
     The argument of `omp_set_num_threads' shall be a positive integer.
587
 
588
_C/C++_:
589
     _Prototype_:  `void omp_set_num_threads(int);'
590
 
591
_Fortran_:
592
     _Interface_:  `subroutine omp_set_num_threads(set)'
593
                   `integer, intent(in) :: set'
594
 
595
_See also_:
596
     *note OMP_NUM_THREADS::, *note omp_get_num_threads::, *note
597
     omp_get_max_threads::
598
 
599
_Reference_:
600
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.2.1.
601
 
602

603
File: libgomp.info,  Node: omp_set_schedule,  Next: omp_init_lock,  Prev: omp_set_num_threads,  Up: Runtime Library Routines
604
 
605
2.19 `omp_set_schedule' - Set the runtime scheduling method
606
===========================================================
607
 
608
_Description_:
609
     Sets the runtime scheduling method. The KIND argument can have the
610
     value `omp_sched_static', `omp_sched_dynamic', `opm_sched_guided'
611
     or `omp_sched_auto'. Except for `omp_sched_auto', the chunk size
612
     is set to the value of MODIFIER if positive or to the default
613
     value if zero or negative.  For `omp_sched_auto' the MODIFIER
614
     argument is ignored.
615
 
616
_C/C++_
617
     _Prototype_:  `int omp_schedule(omp_sched_t * kind, int *modifier);'
618
 
619
_Fortran_:
620
     _Interface_:  `subroutine omp_schedule(kind, modifier)'
621
                   `integer(kind=omp_sched_kind) kind'
622
                   `integer modifier'
623
 
624
_See also_:
625
     *note omp_get_schedule:: *note OMP_SCHEDULE::
626
 
627
_Reference_:
628
     OpenMP specifications v3.0 (http://www.openmp.org/), section
629
     3.2.11.
630
 
631

632
File: libgomp.info,  Node: omp_init_lock,  Next: omp_set_lock,  Prev: omp_set_schedule,  Up: Runtime Library Routines
633
 
634
2.20 `omp_init_lock' - Initialize simple lock
635
=============================================
636
 
637
_Description_:
638
     Initialize a simple lock. After initialization, the lock is in an
639
     unlocked state.
640
 
641
_C/C++_:
642
     _Prototype_:  `void omp_init_lock(omp_lock_t *lock);'
643
 
644
_Fortran_:
645
     _Interface_:  `subroutine omp_init_lock(lock)'
646
                   `integer(omp_lock_kind), intent(out) :: lock'
647
 
648
_See also_:
649
     *note omp_destroy_lock::
650
 
651
_Reference_:
652
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.1.
653
 
654

655
File: libgomp.info,  Node: omp_set_lock,  Next: omp_test_lock,  Prev: omp_init_lock,  Up: Runtime Library Routines
656
 
657
2.21 `omp_set_lock' - Wait for and set simple lock
658
==================================================
659
 
660
_Description_:
661
     Before setting a simple lock, the lock variable must be
662
     initialized by `omp_init_lock'. The calling thread is blocked
663
     until the lock is available. If the lock is already held by the
664
     current thread, a deadlock occurs.
665
 
666
_C/C++_:
667
     _Prototype_:  `void omp_set_lock(omp_lock_t *lock);'
668
 
669
_Fortran_:
670
     _Interface_:  `subroutine omp_set_lock(lock)'
671
                   `integer(omp_lock_kind), intent(out) :: lock'
672
 
673
_See also_:
674
     *note omp_init_lock::, *note omp_test_lock::, *note
675
     omp_unset_lock::
676
 
677
_Reference_:
678
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.3.
679
 
680

681
File: libgomp.info,  Node: omp_test_lock,  Next: omp_unset_lock,  Prev: omp_set_lock,  Up: Runtime Library Routines
682
 
683
2.22 `omp_test_lock' - Test and set simple lock if available
684
============================================================
685
 
686
_Description_:
687
     Before setting a simple lock, the lock variable must be
688
     initialized by `omp_init_lock'. Contrary to `omp_set_lock',
689
     `omp_test_lock' does not block if the lock is not available. This
690
     function returns `true' upon success, `false' otherwise. Here,
691
     `true' and `false' represent their language-specific counterparts.
692
 
693
_C/C++_:
694
     _Prototype_:  `int omp_test_lock(omp_lock_t *lock);'
695
 
696
_Fortran_:
697
     _Interface_:  `subroutine omp_test_lock(lock)'
698
                   `logical(omp_logical_kind) :: omp_test_lock'
699
                   `integer(omp_lock_kind), intent(out) :: lock'
700
 
701
_See also_:
702
     *note omp_init_lock::, *note omp_set_lock::, *note omp_set_lock::
703
 
704
_Reference_:
705
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.5.
706
 
707

708
File: libgomp.info,  Node: omp_unset_lock,  Next: omp_destroy_lock,  Prev: omp_test_lock,  Up: Runtime Library Routines
709
 
710
2.23 `omp_unset_lock' - Unset simple lock
711
=========================================
712
 
713
_Description_:
714
     A simple lock about to be unset must have been locked by
715
     `omp_set_lock' or `omp_test_lock' before. In addition, the lock
716
     must be held by the thread calling `omp_unset_lock'. Then, the
717
     lock becomes unlocked. If one ore more threads attempted to set
718
     the lock before, one of them is chosen to, again, set the lock for
719
     itself.
720
 
721
_C/C++_:
722
     _Prototype_:  `void omp_unset_lock(omp_lock_t *lock);'
723
 
724
_Fortran_:
725
     _Interface_:  `subroutine omp_unset_lock(lock)'
726
                   `integer(omp_lock_kind), intent(out) :: lock'
727
 
728
_See also_:
729
     *note omp_set_lock::, *note omp_test_lock::
730
 
731
_Reference_:
732
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.4.
733
 
734

735
File: libgomp.info,  Node: omp_destroy_lock,  Next: omp_init_nest_lock,  Prev: omp_unset_lock,  Up: Runtime Library Routines
736
 
737
2.24 `omp_destroy_lock' - Destroy simple lock
738
=============================================
739
 
740
_Description_:
741
     Destroy a simple lock. In order to be destroyed, a simple lock
742
     must be in the unlocked state.
743
 
744
_C/C++_:
745
     _Prototype_:  `void omp_destroy_lock(omp_lock_t *);'
746
 
747
_Fortran_:
748
     _Interface_:  `subroutine omp_destroy_lock(lock)'
749
                   `integer(omp_lock_kind), intent(inout) :: lock'
750
 
751
_See also_:
752
     *note omp_init_lock::
753
 
754
_Reference_:
755
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.2.
756
 
757

758
File: libgomp.info,  Node: omp_init_nest_lock,  Next: omp_set_nest_lock,  Prev: omp_destroy_lock,  Up: Runtime Library Routines
759
 
760
2.25 `omp_init_nest_lock' - Initialize nested lock
761
==================================================
762
 
763
_Description_:
764
     Initialize a nested lock. After initialization, the lock is in an
765
     unlocked state and the nesting count is set to zero.
766
 
767
_C/C++_:
768
     _Prototype_:  `void omp_init_nest_lock(omp_nest_lock_t *lock);'
769
 
770
_Fortran_:
771
     _Interface_:  `subroutine omp_init_nest_lock(lock)'
772
                   `integer(omp_nest_lock_kind), intent(out) :: lock'
773
 
774
_See also_:
775
     *note omp_destroy_nest_lock::
776
 
777
_Reference_:
778
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.1.
779
 
780

781
File: libgomp.info,  Node: omp_set_nest_lock,  Next: omp_test_nest_lock,  Prev: omp_init_nest_lock,  Up: Runtime Library Routines
782
 
783
2.26 `omp_set_nest_lock' - Wait for and set simple lock
784
=======================================================
785
 
786
_Description_:
787
     Before setting a nested lock, the lock variable must be
788
     initialized by `omp_init_nest_lock'. The calling thread is blocked
789
     until the lock is available. If the lock is already held by the
790
     current thread, the nesting count for the lock in incremented.
791
 
792
_C/C++_:
793
     _Prototype_:  `void omp_set_nest_lock(omp_nest_lock_t *lock);'
794
 
795
_Fortran_:
796
     _Interface_:  `subroutine omp_set_nest_lock(lock)'
797
                   `integer(omp_nest_lock_kind), intent(out) :: lock'
798
 
799
_See also_:
800
     *note omp_init_nest_lock::, *note omp_unset_nest_lock::
801
 
802
_Reference_:
803
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.3.
804
 
805

806
File: libgomp.info,  Node: omp_test_nest_lock,  Next: omp_unset_nest_lock,  Prev: omp_set_nest_lock,  Up: Runtime Library Routines
807
 
808
2.27 `omp_test_nest_lock' - Test and set nested lock if available
809
=================================================================
810
 
811
_Description_:
812
     Before setting a nested lock, the lock variable must be
813
     initialized by `omp_init_nest_lock'. Contrary to
814
     `omp_set_nest_lock', `omp_test_nest_lock' does not block if the
815
     lock is not available.  If the lock is already held by the current
816
     thread, the new nesting count is returned. Otherwise, the return
817
     value equals zero.
818
 
819
_C/C++_:
820
     _Prototype_:  `int omp_test_nest_lock(omp_nest_lock_t *lock);'
821
 
822
_Fortran_:
823
     _Interface_:  `integer function omp_test_nest_lock(lock)'
824
                   `integer(omp_integer_kind) :: omp_test_nest_lock'
825
                   `integer(omp_nest_lock_kind), intent(inout) :: lock'
826
 
827
_See also_:
828
     *note omp_init_lock::, *note omp_set_lock::, *note omp_set_lock::
829
 
830
_Reference_:
831
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.5.
832
 
833

834
File: libgomp.info,  Node: omp_unset_nest_lock,  Next: omp_destroy_nest_lock,  Prev: omp_test_nest_lock,  Up: Runtime Library Routines
835
 
836
2.28 `omp_unset_nest_lock' - Unset nested lock
837
==============================================
838
 
839
_Description_:
840
     A nested lock about to be unset must have been locked by
841
     `omp_set_nested_lock' or `omp_test_nested_lock' before. In
842
     addition, the lock must be held by the thread calling
843
     `omp_unset_nested_lock'. If the nesting count drops to zero, the
844
     lock becomes unlocked. If one ore more threads attempted to set
845
     the lock before, one of them is chosen to, again, set the lock for
846
     itself.
847
 
848
_C/C++_:
849
     _Prototype_:  `void omp_unset_nest_lock(omp_nest_lock_t *lock);'
850
 
851
_Fortran_:
852
     _Interface_:  `subroutine omp_unset_nest_lock(lock)'
853
                   `integer(omp_nest_lock_kind), intent(out) :: lock'
854
 
855
_See also_:
856
     *note omp_set_nest_lock::
857
 
858
_Reference_:
859
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.4.
860
 
861

862
File: libgomp.info,  Node: omp_destroy_nest_lock,  Next: omp_get_wtick,  Prev: omp_unset_nest_lock,  Up: Runtime Library Routines
863
 
864
2.29 `omp_destroy_nest_lock' - Destroy nested lock
865
==================================================
866
 
867
_Description_:
868
     Destroy a nested lock. In order to be destroyed, a nested lock
869
     must be in the unlocked state and its nesting count must equal
870
     zero.
871
 
872
_C/C++_:
873
     _Prototype_:  `void omp_destroy_nest_lock(omp_nest_lock_t *);'
874
 
875
_Fortran_:
876
     _Interface_:  `subroutine omp_destroy_nest_lock(lock)'
877
                   `integer(omp_nest_lock_kind), intent(inout) :: lock'
878
 
879
_See also_:
880
     *note omp_init_lock::
881
 
882
_Reference_:
883
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.3.2.
884
 
885

886
File: libgomp.info,  Node: omp_get_wtick,  Next: omp_get_wtime,  Prev: omp_destroy_nest_lock,  Up: Runtime Library Routines
887
 
888
2.30 `omp_get_wtick' - Get timer precision
889
==========================================
890
 
891
_Description_:
892
     Gets the timer precision, i.e., the number of seconds between two
893
     successive clock ticks.
894
 
895
_C/C++_:
896
     _Prototype_:  `double omp_get_wtick();'
897
 
898
_Fortran_:
899
     _Interface_:  `double precision function omp_get_wtick()'
900
 
901
_See also_:
902
     *note omp_get_wtime::
903
 
904
_Reference_:
905
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.4.2.
906
 
907

908
File: libgomp.info,  Node: omp_get_wtime,  Prev: omp_get_wtick,  Up: Runtime Library Routines
909
 
910
2.31 `omp_get_wtime' - Elapsed wall clock time
911
==============================================
912
 
913
_Description_:
914
     Elapsed wall clock time in seconds. The time is measured per
915
     thread, no guarantee can bee made that two distinct threads
916
     measure the same time.  Time is measured from some "time in the
917
     past". On POSIX compliant systems the seconds since the Epoch
918
     (00:00:00 UTC, January 1, 1970) are returned.
919
 
920
_C/C++_:
921
     _Prototype_:  `double omp_get_wtime();'
922
 
923
_Fortran_:
924
     _Interface_:  `double precision function omp_get_wtime()'
925
 
926
_See also_:
927
     *note omp_get_wtick::
928
 
929
_Reference_:
930
     OpenMP specifications v3.0 (http://www.openmp.org/), section 3.4.1.
931
 
932

933
File: libgomp.info,  Node: Environment Variables,  Next: The libgomp ABI,  Prev: Runtime Library Routines,  Up: Top
934
 
935
3 Environment Variables
936
***********************
937
 
938
The variables `OMP_DYNAMIC', `OMP_MAX_ACTIVE_LEVELS', `OMP_NESTED',
939
`OMP_NUM_THREADS', `OMP_SCHEDULE', `OMP_STACKSIZE',`OMP_THREAD_LIMIT'
940
and `OMP_WAIT_POLICY' are defined by section 4 of the OpenMP
941
specifications in version 3.0, while `GOMP_CPU_AFFINITY' and
942
`GOMP_STACKSIZE' are GNU extensions.
943
 
944
* Menu:
945
 
946
* OMP_DYNAMIC::           Dynamic adjustment of threads
947
* OMP_MAX_ACTIVE_LEVELS:: Set the maximal number of nested parallel regions
948
* OMP_NESTED::            Nested parallel regions
949
* OMP_NUM_THREADS::       Specifies the number of threads to use
950
* OMP_STACKSIZE::         Set default thread stack size
951
* OMP_SCHEDULE::          How threads are scheduled
952
* OMP_THREAD_LIMIT::      Set the maximal number of threads
953
* OMP_WAIT_POLICY::       How waiting threads are handled
954
* GOMP_CPU_AFFINITY::     Bind threads to specific CPUs
955
* GOMP_STACKSIZE::        Set default thread stack size
956
 
957

958
File: libgomp.info,  Node: OMP_DYNAMIC,  Next: OMP_MAX_ACTIVE_LEVELS,  Up: Environment Variables
959
 
960
3.1 `OMP_DYNAMIC' - Dynamic adjustment of threads
961
=================================================
962
 
963
_Description_:
964
     Enable or disable the dynamic adjustment of the number of threads
965
     within a team. The value of this environment variable shall be
966
     `TRUE' or `FALSE'. If undefined, dynamic adjustment is disabled by
967
     default.
968
 
969
_See also_:
970
     *note omp_set_dynamic::
971
 
972
_Reference_:
973
     OpenMP specifications v3.0 (http://www.openmp.org/), section 4.3
974
 
975

976
File: libgomp.info,  Node: OMP_MAX_ACTIVE_LEVELS,  Next: OMP_NESTED,  Prev: OMP_DYNAMIC,  Up: Environment Variables
977
 
978
3.2 `OMP_MAX_ACTIVE_LEVELS' - Set the maximal number of nested parallel regions
979
===============================================================================
980
 
981
_Description_:
982
     Specifies the initial value for the maximal number of nested
983
     parallel regions. The value of this variable shall be positive
984
     integer.  If undefined, the number of active levels is unlimited.
985
 
986
_See also_:
987
     *note omp_set_max_active_levels::
988
 
989
_Reference_:
990
     OpenMP specifications v3.0 (http://www.openmp.org/), section 4.7
991
 
992

993
File: libgomp.info,  Node: OMP_NESTED,  Next: OMP_NUM_THREADS,  Prev: OMP_MAX_ACTIVE_LEVELS,  Up: Environment Variables
994
 
995
3.3 `OMP_NESTED' - Nested parallel regions
996
==========================================
997
 
998
_Description_:
999
     Enable or disable nested parallel regions, i.e., whether team
1000
     members are allowed to create new teams. The value of this
1001
     environment variable shall be `TRUE' or `FALSE'. If undefined,
1002
     nested parallel regions are disabled by default.
1003
 
1004
_See also_:
1005
     *note omp_set_nested::
1006
 
1007
_Reference_:
1008
     OpenMP specifications v3.0 (http://www.openmp.org/), section 4.4
1009
 
1010

1011
File: libgomp.info,  Node: OMP_NUM_THREADS,  Next: OMP_STACKSIZE,  Prev: OMP_NESTED,  Up: Environment Variables
1012
 
1013
3.4 `OMP_NUM_THREADS' - Specifies the number of threads to use
1014
==============================================================
1015
 
1016
_Description_:
1017
     Specifies the default number of threads to use in parallel
1018
     regions. The value of this variable shall be positive integer. If
1019
     undefined one thread per CPU online is used.
1020
 
1021
_See also_:
1022
     *note omp_set_num_threads::
1023
 
1024
_Reference_:
1025
     OpenMP specifications v3.0 (http://www.openmp.org/), section 4.2
1026
 
1027

1028
File: libgomp.info,  Node: OMP_SCHEDULE,  Next: OMP_THREAD_LIMIT,  Prev: OMP_STACKSIZE,  Up: Environment Variables
1029
 
1030
3.5 `OMP_SCHEDULE' - How threads are scheduled
1031
==============================================
1032
 
1033
_Description_:
1034
     Allows to specify `schedule type' and `chunk size'.  The value of
1035
     the variable shall have the form: `type[,chunk]' where `type' is
1036
     one of `static', `dynamic', `guided' or `auto' The optional
1037
     `chunk' size shall be a positive integer. If undefined, dynamic
1038
     scheduling and a chunk size of 1 is used.
1039
 
1040
_See also_:
1041
     *note omp_set_schedule::
1042
 
1043
_Reference_:
1044
     OpenMP specifications v3.0 (http://www.openmp.org/), sections
1045
     2.5.1 and 4.1
1046
 
1047

1048
File: libgomp.info,  Node: OMP_STACKSIZE,  Next: OMP_SCHEDULE,  Prev: OMP_NUM_THREADS,  Up: Environment Variables
1049
 
1050
3.6 `OMP_STACKSIZE' - Set default thread stack size
1051
===================================================
1052
 
1053
_Description_:
1054
     Set the default thread stack size in kilobytes, unless the number
1055
     is suffixed by `B', `K', `M' or `G', in which case the size is,
1056
     respectively, in bytes, kilobytes, megabytes or gigabytes. This is
1057
     different from `pthread_attr_setstacksize' which gets the number
1058
     of bytes as an argument. If the stacksize can not be set due to
1059
     system constraints, an error is reported and the initial stacksize
1060
     is left unchanged. If undefined, the stack size is system
1061
     dependent.
1062
 
1063
_Reference_:
1064
     OpenMP specifications v3.0 (http://www.openmp.org/), sections 4.5
1065
 
1066

1067
File: libgomp.info,  Node: OMP_THREAD_LIMIT,  Next: OMP_WAIT_POLICY,  Prev: OMP_SCHEDULE,  Up: Environment Variables
1068
 
1069
3.7 `OMP_THREAD_LIMIT' - Set the maximal number of threads
1070
==========================================================
1071
 
1072
_Description_:
1073
     Specifies the number of threads to use for the whole program. The
1074
     value of this variable shall be positive integer. If undefined,
1075
     the number of threads is not limited.
1076
 
1077
_See also_:
1078
     *note OMP_NUM_THREADS:: *note omp_get_thread_limit::
1079
 
1080
_Reference_:
1081
     OpenMP specifications v3.0 (http://www.openmp.org/), section 4.8
1082
 
1083

1084
File: libgomp.info,  Node: OMP_WAIT_POLICY,  Next: GOMP_CPU_AFFINITY,  Prev: OMP_THREAD_LIMIT,  Up: Environment Variables
1085
 
1086
3.8 `OMP_WAIT_POLICY' - How waiting threads are handled
1087
=======================================================
1088
 
1089
_Description_:
1090
     Specifies whether waiting threads should be active or passive. If
1091
     the value is `PASSIVE', waiting threads should not consume CPU
1092
     power while waiting; while the value is `ACTIVE' specifies that
1093
     they should.
1094
 
1095
_Reference_:
1096
     OpenMP specifications v3.0 (http://www.openmp.org/), sections 4.6
1097
 
1098

1099
File: libgomp.info,  Node: GOMP_CPU_AFFINITY,  Next: GOMP_STACKSIZE,  Prev: OMP_WAIT_POLICY,  Up: Environment Variables
1100
 
1101
3.9 `GOMP_CPU_AFFINITY' - Bind threads to specific CPUs
1102
=======================================================
1103
 
1104
_Description_:
1105
     Binds threads to specific CPUs. The variable should contain a
1106
     space- or comma-separated list of CPUs. This list may contain
1107
     different kind of entries: either single CPU numbers in any order,
1108
     a range of CPUs (M-N) or a range with some stride (M-N:S). CPU
1109
     numbers are zero based. For example, `GOMP_CPU_AFFINITY="0 3 1-2
1110
     4-15:2"' will bind the initial thread to CPU 0, the second to CPU
1111
     3, the third to CPU 1, the fourth to CPU 2, the fifth to CPU 4,
1112
     the sixth through tenth to CPUs 6, 8, 10, 12, and 14 respectively
1113
     and then start assigning back from the beginning of the list.
1114
     `GOMP_CPU_AFFINITY=0' binds all threads to CPU 0.
1115
 
1116
     There is no GNU OpenMP library routine to determine whether a CPU
1117
     affinity specification is in effect. As a workaround,
1118
     language-specific library functions, e.g., `getenv' in C or
1119
     `GET_ENVIRONMENT_VARIABLE' in Fortran, may be used to query the
1120
     setting of the `GOMP_CPU_AFFINITY' environment variable. A defined
1121
     CPU affinity on startup cannot be changed or disabled during the
1122
     runtime of the application.
1123
 
1124
     If this environment variable is omitted, the host system will
1125
     handle the assignment of threads to CPUs.
1126
 
1127

1128
File: libgomp.info,  Node: GOMP_STACKSIZE,  Prev: GOMP_CPU_AFFINITY,  Up: Environment Variables
1129
 
1130
3.10 `GOMP_STACKSIZE' - Set default thread stack size
1131
=====================================================
1132
 
1133
_Description_:
1134
     Set the default thread stack size in kilobytes. This is different
1135
     from `pthread_attr_setstacksize' which gets the number of bytes as
1136
     an argument. If the stacksize can not be set due to system
1137
     constraints, an error is reported and the initial stacksize is
1138
     left unchanged. If undefined, the stack size is system dependent.
1139
 
1140
_See also_:
1141
     *note GOMP_STACKSIZE::
1142
 
1143
_Reference_:
1144
     GCC Patches Mailinglist
1145
     (http://gcc.gnu.org/ml/gcc-patches/2006-06/msg00493.html), GCC
1146
     Patches Mailinglist
1147
     (http://gcc.gnu.org/ml/gcc-patches/2006-06/msg00496.html)
1148
 
1149

1150
File: libgomp.info,  Node: The libgomp ABI,  Next: Reporting Bugs,  Prev: Environment Variables,  Up: Top
1151
 
1152
4 The libgomp ABI
1153
*****************
1154
 
1155
The following sections present notes on the external ABI as presented
1156
by libgomp. Only maintainers should need them.
1157
 
1158
* Menu:
1159
 
1160
* Implementing MASTER construct::
1161
* Implementing CRITICAL construct::
1162
* Implementing ATOMIC construct::
1163
* Implementing FLUSH construct::
1164
* Implementing BARRIER construct::
1165
* Implementing THREADPRIVATE construct::
1166
* Implementing PRIVATE clause::
1167
* Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses::
1168
* Implementing REDUCTION clause::
1169
* Implementing PARALLEL construct::
1170
* Implementing FOR construct::
1171
* Implementing ORDERED construct::
1172
* Implementing SECTIONS construct::
1173
* Implementing SINGLE construct::
1174
 
1175

1176
File: libgomp.info,  Node: Implementing MASTER construct,  Next: Implementing CRITICAL construct,  Up: The libgomp ABI
1177
 
1178
4.1 Implementing MASTER construct
1179
=================================
1180
 
1181
     if (omp_get_thread_num () == 0)
1182
       block
1183
 
1184
   Alternately, we generate two copies of the parallel subfunction and
1185
only include this in the version run by the master thread.  Surely
1186
that's not worthwhile though...
1187
 
1188

1189
File: libgomp.info,  Node: Implementing CRITICAL construct,  Next: Implementing ATOMIC construct,  Prev: Implementing MASTER construct,  Up: The libgomp ABI
1190
 
1191
4.2 Implementing CRITICAL construct
1192
===================================
1193
 
1194
Without a specified name,
1195
 
1196
       void GOMP_critical_start (void);
1197
       void GOMP_critical_end (void);
1198
 
1199
   so that we don't get COPY relocations from libgomp to the main
1200
application.
1201
 
1202
   With a specified name, use omp_set_lock and omp_unset_lock with name
1203
being transformed into a variable declared like
1204
 
1205
       omp_lock_t gomp_critical_user_ __attribute__((common))
1206
 
1207
   Ideally the ABI would specify that all zero is a valid unlocked
1208
state, and so we wouldn't actually need to initialize this at startup.
1209
 
1210

1211
File: libgomp.info,  Node: Implementing ATOMIC construct,  Next: Implementing FLUSH construct,  Prev: Implementing CRITICAL construct,  Up: The libgomp ABI
1212
 
1213
4.3 Implementing ATOMIC construct
1214
=================================
1215
 
1216
The target should implement the `__sync' builtins.
1217
 
1218
   Failing that we could add
1219
 
1220
       void GOMP_atomic_enter (void)
1221
       void GOMP_atomic_exit (void)
1222
 
1223
   which reuses the regular lock code, but with yet another lock object
1224
private to the library.
1225
 
1226

1227
File: libgomp.info,  Node: Implementing FLUSH construct,  Next: Implementing BARRIER construct,  Prev: Implementing ATOMIC construct,  Up: The libgomp ABI
1228
 
1229
4.4 Implementing FLUSH construct
1230
================================
1231
 
1232
Expands to the `__sync_synchronize' builtin.
1233
 
1234

1235
File: libgomp.info,  Node: Implementing BARRIER construct,  Next: Implementing THREADPRIVATE construct,  Prev: Implementing FLUSH construct,  Up: The libgomp ABI
1236
 
1237
4.5 Implementing BARRIER construct
1238
==================================
1239
 
1240
       void GOMP_barrier (void)
1241
 
1242

1243
File: libgomp.info,  Node: Implementing THREADPRIVATE construct,  Next: Implementing PRIVATE clause,  Prev: Implementing BARRIER construct,  Up: The libgomp ABI
1244
 
1245
4.6 Implementing THREADPRIVATE construct
1246
========================================
1247
 
1248
In _most_ cases we can map this directly to `__thread'.  Except that
1249
OMP allows constructors for C++ objects.  We can either refuse to
1250
support this (how often is it used?) or we can implement something akin
1251
to .ctors.
1252
 
1253
   Even more ideally, this ctor feature is handled by extensions to the
1254
main pthreads library.  Failing that, we can have a set of entry points
1255
to register ctor functions to be called.
1256
 
1257

1258
File: libgomp.info,  Node: Implementing PRIVATE clause,  Next: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses,  Prev: Implementing THREADPRIVATE construct,  Up: The libgomp ABI
1259
 
1260
4.7 Implementing PRIVATE clause
1261
===============================
1262
 
1263
In association with a PARALLEL, or within the lexical extent of a
1264
PARALLEL block, the variable becomes a local variable in the parallel
1265
subfunction.
1266
 
1267
   In association with FOR or SECTIONS blocks, create a new automatic
1268
variable within the current function.  This preserves the semantic of
1269
new variable creation.
1270
 
1271

1272
File: libgomp.info,  Node: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses,  Next: Implementing REDUCTION clause,  Prev: Implementing PRIVATE clause,  Up: The libgomp ABI
1273
 
1274
4.8 Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses
1275
========================================================================
1276
 
1277
Seems simple enough for PARALLEL blocks.  Create a private struct for
1278
communicating between parent and subfunction.  In the parent, copy in
1279
values for scalar and "small" structs; copy in addresses for others
1280
TREE_ADDRESSABLE types.  In the subfunction, copy the value into the
1281
local variable.
1282
 
1283
   Not clear at all what to do with bare FOR or SECTION blocks.  The
1284
only thing I can figure is that we do something like
1285
 
1286
     #pragma omp for firstprivate(x) lastprivate(y)
1287
     for (int i = 0; i < n; ++i)
1288
       body;
1289
 
1290
   which becomes
1291
 
1292
     {
1293
       int x = x, y;
1294
 
1295
       // for stuff
1296
 
1297
       if (i == n)
1298
         y = y;
1299
     }
1300
 
1301
   where the "x=x" and "y=y" assignments actually have different uids
1302
for the two variables, i.e. not something you could write directly in
1303
C.  Presumably this only makes sense if the "outer" x and y are global
1304
variables.
1305
 
1306
   COPYPRIVATE would work the same way, except the structure broadcast
1307
would have to happen via SINGLE machinery instead.
1308
 
1309

1310
File: libgomp.info,  Node: Implementing REDUCTION clause,  Next: Implementing PARALLEL construct,  Prev: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses,  Up: The libgomp ABI
1311
 
1312
4.9 Implementing REDUCTION clause
1313
=================================
1314
 
1315
The private struct mentioned in the previous section should have a
1316
pointer to an array of the type of the variable, indexed by the
1317
thread's TEAM_ID.  The thread stores its final value into the array,
1318
and after the barrier the master thread iterates over the array to
1319
collect the values.
1320
 
1321

1322
File: libgomp.info,  Node: Implementing PARALLEL construct,  Next: Implementing FOR construct,  Prev: Implementing REDUCTION clause,  Up: The libgomp ABI
1323
 
1324
4.10 Implementing PARALLEL construct
1325
====================================
1326
 
1327
       #pragma omp parallel
1328
       {
1329
         body;
1330
       }
1331
 
1332
   becomes
1333
 
1334
       void subfunction (void *data)
1335
       {
1336
         use data;
1337
         body;
1338
       }
1339
 
1340
       setup data;
1341
       GOMP_parallel_start (subfunction, &data, num_threads);
1342
       subfunction (&data);
1343
       GOMP_parallel_end ();
1344
 
1345
       void GOMP_parallel_start (void (*fn)(void *), void *data, unsigned num_threads)
1346
 
1347
   The FN argument is the subfunction to be run in parallel.
1348
 
1349
   The DATA argument is a pointer to a structure used to communicate
1350
data in and out of the subfunction, as discussed above with respect to
1351
FIRSTPRIVATE et al.
1352
 
1353
   The NUM_THREADS argument is 1 if an IF clause is present and false,
1354
or the value of the NUM_THREADS clause, if present, or 0.
1355
 
1356
   The function needs to create the appropriate number of threads
1357
and/or launch them from the dock.  It needs to create the team
1358
structure and assign team ids.
1359
 
1360
       void GOMP_parallel_end (void)
1361
 
1362
   Tears down the team and returns us to the previous
1363
`omp_in_parallel()' state.
1364
 
1365

1366
File: libgomp.info,  Node: Implementing FOR construct,  Next: Implementing ORDERED construct,  Prev: Implementing PARALLEL construct,  Up: The libgomp ABI
1367
 
1368
4.11 Implementing FOR construct
1369
===============================
1370
 
1371
       #pragma omp parallel for
1372
       for (i = lb; i <= ub; i++)
1373
         body;
1374
 
1375
   becomes
1376
 
1377
       void subfunction (void *data)
1378
       {
1379
         long _s0, _e0;
1380
         while (GOMP_loop_static_next (&_s0, &_e0))
1381
         {
1382
           long _e1 = _e0, i;
1383
           for (i = _s0; i < _e1; i++)
1384
             body;
1385
         }
1386
         GOMP_loop_end_nowait ();
1387
       }
1388
 
1389
       GOMP_parallel_loop_static (subfunction, NULL, 0, lb, ub+1, 1, 0);
1390
       subfunction (NULL);
1391
       GOMP_parallel_end ();
1392
 
1393
       #pragma omp for schedule(runtime)
1394
       for (i = 0; i < n; i++)
1395
         body;
1396
 
1397
   becomes
1398
 
1399
       {
1400
         long i, _s0, _e0;
1401
         if (GOMP_loop_runtime_start (0, n, 1, &_s0, &_e0))
1402
           do {
1403
             long _e1 = _e0;
1404
             for (i = _s0, i < _e0; i++)
1405
               body;
1406
           } while (GOMP_loop_runtime_next (&_s0, _&e0));
1407
         GOMP_loop_end ();
1408
       }
1409
 
1410
   Note that while it looks like there is trickyness to propagating a
1411
non-constant STEP, there isn't really.  We're explicitly allowed to
1412
evaluate it as many times as we want, and any variables involved should
1413
automatically be handled as PRIVATE or SHARED like any other variables.
1414
So the expression should remain evaluable in the subfunction.  We can
1415
also pull it into a local variable if we like, but since its supposed
1416
to remain unchanged, we can also not if we like.
1417
 
1418
   If we have SCHEDULE(STATIC), and no ORDERED, then we ought to be
1419
able to get away with no work-sharing context at all, since we can
1420
simply perform the arithmetic directly in each thread to divide up the
1421
iterations.  Which would mean that we wouldn't need to call any of
1422
these routines.
1423
 
1424
   There are separate routines for handling loops with an ORDERED
1425
clause.  Bookkeeping for that is non-trivial...
1426
 
1427

1428
File: libgomp.info,  Node: Implementing ORDERED construct,  Next: Implementing SECTIONS construct,  Prev: Implementing FOR construct,  Up: The libgomp ABI
1429
 
1430
4.12 Implementing ORDERED construct
1431
===================================
1432
 
1433
       void GOMP_ordered_start (void)
1434
       void GOMP_ordered_end (void)
1435
 
1436

1437
File: libgomp.info,  Node: Implementing SECTIONS construct,  Next: Implementing SINGLE construct,  Prev: Implementing ORDERED construct,  Up: The libgomp ABI
1438
 
1439
4.13 Implementing SECTIONS construct
1440
====================================
1441
 
1442
A block as
1443
 
1444
       #pragma omp sections
1445
       {
1446
         #pragma omp section
1447
         stmt1;
1448
         #pragma omp section
1449
         stmt2;
1450
         #pragma omp section
1451
         stmt3;
1452
       }
1453
 
1454
   becomes
1455
 
1456
       for (i = GOMP_sections_start (3); i != 0; i = GOMP_sections_next ())
1457
         switch (i)
1458
           {
1459
           case 1:
1460
             stmt1;
1461
             break;
1462
           case 2:
1463
             stmt2;
1464
             break;
1465
           case 3:
1466
             stmt3;
1467
             break;
1468
           }
1469
       GOMP_barrier ();
1470
 
1471

1472
File: libgomp.info,  Node: Implementing SINGLE construct,  Prev: Implementing SECTIONS construct,  Up: The libgomp ABI
1473
 
1474
4.14 Implementing SINGLE construct
1475
==================================
1476
 
1477
A block like
1478
 
1479
       #pragma omp single
1480
       {
1481
         body;
1482
       }
1483
 
1484
   becomes
1485
 
1486
       if (GOMP_single_start ())
1487
         body;
1488
       GOMP_barrier ();
1489
 
1490
   while
1491
 
1492
       #pragma omp single copyprivate(x)
1493
         body;
1494
 
1495
   becomes
1496
 
1497
       datap = GOMP_single_copy_start ();
1498
       if (datap == NULL)
1499
         {
1500
           body;
1501
           data.x = x;
1502
           GOMP_single_copy_end (&data);
1503
         }
1504
       else
1505
         x = datap->x;
1506
       GOMP_barrier ();
1507
 
1508

1509
File: libgomp.info,  Node: Reporting Bugs,  Next: Copying,  Prev: The libgomp ABI,  Up: Top
1510
 
1511
5 Reporting Bugs
1512
****************
1513
 
1514
Bugs in the GNU OpenMP implementation should be reported via bugzilla
1515
(http://gcc.gnu.org/bugzilla/). In all cases, please add "openmp" to
1516
the keywords field in the bug report.
1517
 
1518

1519
File: libgomp.info,  Node: Copying,  Next: GNU Free Documentation License,  Prev: Reporting Bugs,  Up: Top
1520
 
1521
GNU GENERAL PUBLIC LICENSE
1522
**************************
1523
 
1524
                         Version 2, June 1991
1525
 
1526
     Copyright (C) 1989, 1991 Free Software Foundation, Inc.
1527
     51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
1528
 
1529
     Everyone is permitted to copy and distribute verbatim copies
1530
     of this license document, but changing it is not allowed.
1531
 
1532
Preamble
1533
========
1534
 
1535
The licenses for most software are designed to take away your freedom
1536
to share and change it.  By contrast, the GNU General Public License is
1537
intended to guarantee your freedom to share and change free
1538
software--to make sure the software is free for all its users.  This
1539
General Public License applies to most of the Free Software
1540
Foundation's software and to any other program whose authors commit to
1541
using it.  (Some other Free Software Foundation software is covered by
1542
the GNU Library General Public License instead.)  You can apply it to
1543
your programs, too.
1544
 
1545
   When we speak of free software, we are referring to freedom, not
1546
price.  Our General Public Licenses are designed to make sure that you
1547
have the freedom to distribute copies of free software (and charge for
1548
this service if you wish), that you receive source code or can get it
1549
if you want it, that you can change the software or use pieces of it in
1550
new free programs; and that you know you can do these things.
1551
 
1552
   To protect your rights, we need to make restrictions that forbid
1553
anyone to deny you these rights or to ask you to surrender the rights.
1554
These restrictions translate to certain responsibilities for you if you
1555
distribute copies of the software, or if you modify it.
1556
 
1557
   For example, if you distribute copies of such a program, whether
1558
gratis or for a fee, you must give the recipients all the rights that
1559
you have.  You must make sure that they, too, receive or can get the
1560
source code.  And you must show them these terms so they know their
1561
rights.
1562
 
1563
   We protect your rights with two steps: (1) copyright the software,
1564
and (2) offer you this license which gives you legal permission to copy,
1565
distribute and/or modify the software.
1566
 
1567
   Also, for each author's protection and ours, we want to make certain
1568
that everyone understands that there is no warranty for this free
1569
software.  If the software is modified by someone else and passed on, we
1570
want its recipients to know that what they have is not the original, so
1571
that any problems introduced by others will not reflect on the original
1572
authors' reputations.
1573
 
1574
   Finally, any free program is threatened constantly by software
1575
patents.  We wish to avoid the danger that redistributors of a free
1576
program will individually obtain patent licenses, in effect making the
1577
program proprietary.  To prevent this, we have made it clear that any
1578
patent must be licensed for everyone's free use or not licensed at all.
1579
 
1580
   The precise terms and conditions for copying, distribution and
1581
modification follow.
1582
 
1583
    TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
1584
  0. This License applies to any program or other work which contains a
1585
     notice placed by the copyright holder saying it may be distributed
1586
     under the terms of this General Public License.  The "Program",
1587
     below, refers to any such program or work, and a "work based on
1588
     the Program" means either the Program or any derivative work under
1589
     copyright law: that is to say, a work containing the Program or a
1590
     portion of it, either verbatim or with modifications and/or
1591
     translated into another language.  (Hereinafter, translation is
1592
     included without limitation in the term "modification".)  Each
1593
     licensee is addressed as "you".
1594
 
1595
     Activities other than copying, distribution and modification are
1596
     not covered by this License; they are outside its scope.  The act
1597
     of running the Program is not restricted, and the output from the
1598
     Program is covered only if its contents constitute a work based on
1599
     the Program (independent of having been made by running the
1600
     Program).  Whether that is true depends on what the Program does.
1601
 
1602
  1. You may copy and distribute verbatim copies of the Program's
1603
     source code as you receive it, in any medium, provided that you
1604
     conspicuously and appropriately publish on each copy an appropriate
1605
     copyright notice and disclaimer of warranty; keep intact all the
1606
     notices that refer to this License and to the absence of any
1607
     warranty; and give any other recipients of the Program a copy of
1608
     this License along with the Program.
1609
 
1610
     You may charge a fee for the physical act of transferring a copy,
1611
     and you may at your option offer warranty protection in exchange
1612
     for a fee.
1613
 
1614
  2. You may modify your copy or copies of the Program or any portion
1615
     of it, thus forming a work based on the Program, and copy and
1616
     distribute such modifications or work under the terms of Section 1
1617
     above, provided that you also meet all of these conditions:
1618
 
1619
       a. You must cause the modified files to carry prominent notices
1620
          stating that you changed the files and the date of any change.
1621
 
1622
       b. You must cause any work that you distribute or publish, that
1623
          in whole or in part contains or is derived from the Program
1624
          or any part thereof, to be licensed as a whole at no charge
1625
          to all third parties under the terms of this License.
1626
 
1627
       c. If the modified program normally reads commands interactively
1628
          when run, you must cause it, when started running for such
1629
          interactive use in the most ordinary way, to print or display
1630
          an announcement including an appropriate copyright notice and
1631
          a notice that there is no warranty (or else, saying that you
1632
          provide a warranty) and that users may redistribute the
1633
          program under these conditions, and telling the user how to
1634
          view a copy of this License.  (Exception: if the Program
1635
          itself is interactive but does not normally print such an
1636
          announcement, your work based on the Program is not required
1637
          to print an announcement.)
1638
 
1639
     These requirements apply to the modified work as a whole.  If
1640
     identifiable sections of that work are not derived from the
1641
     Program, and can be reasonably considered independent and separate
1642
     works in themselves, then this License, and its terms, do not
1643
     apply to those sections when you distribute them as separate
1644
     works.  But when you distribute the same sections as part of a
1645
     whole which is a work based on the Program, the distribution of
1646
     the whole must be on the terms of this License, whose permissions
1647
     for other licensees extend to the entire whole, and thus to each
1648
     and every part regardless of who wrote it.
1649
 
1650
     Thus, it is not the intent of this section to claim rights or
1651
     contest your rights to work written entirely by you; rather, the
1652
     intent is to exercise the right to control the distribution of
1653
     derivative or collective works based on the Program.
1654
 
1655
     In addition, mere aggregation of another work not based on the
1656
     Program with the Program (or with a work based on the Program) on
1657
     a volume of a storage or distribution medium does not bring the
1658
     other work under the scope of this License.
1659
 
1660
  3. You may copy and distribute the Program (or a work based on it,
1661
     under Section 2) in object code or executable form under the terms
1662
     of Sections 1 and 2 above provided that you also do one of the
1663
     following:
1664
 
1665
       a. Accompany it with the complete corresponding machine-readable
1666
          source code, which must be distributed under the terms of
1667
          Sections 1 and 2 above on a medium customarily used for
1668
          software interchange; or,
1669
 
1670
       b. Accompany it with a written offer, valid for at least three
1671
          years, to give any third party, for a charge no more than your
1672
          cost of physically performing source distribution, a complete
1673
          machine-readable copy of the corresponding source code, to be
1674
          distributed under the terms of Sections 1 and 2 above on a
1675
          medium customarily used for software interchange; or,
1676
 
1677
       c. Accompany it with the information you received as to the offer
1678
          to distribute corresponding source code.  (This alternative is
1679
          allowed only for noncommercial distribution and only if you
1680
          received the program in object code or executable form with
1681
          such an offer, in accord with Subsection b above.)
1682
 
1683
     The source code for a work means the preferred form of the work for
1684
     making modifications to it.  For an executable work, complete
1685
     source code means all the source code for all modules it contains,
1686
     plus any associated interface definition files, plus the scripts
1687
     used to control compilation and installation of the executable.
1688
     However, as a special exception, the source code distributed need
1689
     not include anything that is normally distributed (in either
1690
     source or binary form) with the major components (compiler,
1691
     kernel, and so on) of the operating system on which the executable
1692
     runs, unless that component itself accompanies the executable.
1693
 
1694
     If distribution of executable or object code is made by offering
1695
     access to copy from a designated place, then offering equivalent
1696
     access to copy the source code from the same place counts as
1697
     distribution of the source code, even though third parties are not
1698
     compelled to copy the source along with the object code.
1699
 
1700
  4. You may not copy, modify, sublicense, or distribute the Program
1701
     except as expressly provided under this License.  Any attempt
1702
     otherwise to copy, modify, sublicense or distribute the Program is
1703
     void, and will automatically terminate your rights under this
1704
     License.  However, parties who have received copies, or rights,
1705
     from you under this License will not have their licenses
1706
     terminated so long as such parties remain in full compliance.
1707
 
1708
  5. You are not required to accept this License, since you have not
1709
     signed it.  However, nothing else grants you permission to modify
1710
     or distribute the Program or its derivative works.  These actions
1711
     are prohibited by law if you do not accept this License.
1712
     Therefore, by modifying or distributing the Program (or any work
1713
     based on the Program), you indicate your acceptance of this
1714
     License to do so, and all its terms and conditions for copying,
1715
     distributing or modifying the Program or works based on it.
1716
 
1717
  6. Each time you redistribute the Program (or any work based on the
1718
     Program), the recipient automatically receives a license from the
1719
     original licensor to copy, distribute or modify the Program
1720
     subject to these terms and conditions.  You may not impose any
1721
     further restrictions on the recipients' exercise of the rights
1722
     granted herein.  You are not responsible for enforcing compliance
1723
     by third parties to this License.
1724
 
1725
  7. If, as a consequence of a court judgment or allegation of patent
1726
     infringement or for any other reason (not limited to patent
1727
     issues), conditions are imposed on you (whether by court order,
1728
     agreement or otherwise) that contradict the conditions of this
1729
     License, they do not excuse you from the conditions of this
1730
     License.  If you cannot distribute so as to satisfy simultaneously
1731
     your obligations under this License and any other pertinent
1732
     obligations, then as a consequence you may not distribute the
1733
     Program at all.  For example, if a patent license would not permit
1734
     royalty-free redistribution of the Program by all those who
1735
     receive copies directly or indirectly through you, then the only
1736
     way you could satisfy both it and this License would be to refrain
1737
     entirely from distribution of the Program.
1738
 
1739
     If any portion of this section is held invalid or unenforceable
1740
     under any particular circumstance, the balance of the section is
1741
     intended to apply and the section as a whole is intended to apply
1742
     in other circumstances.
1743
 
1744
     It is not the purpose of this section to induce you to infringe any
1745
     patents or other property right claims or to contest validity of
1746
     any such claims; this section has the sole purpose of protecting
1747
     the integrity of the free software distribution system, which is
1748
     implemented by public license practices.  Many people have made
1749
     generous contributions to the wide range of software distributed
1750
     through that system in reliance on consistent application of that
1751
     system; it is up to the author/donor to decide if he or she is
1752
     willing to distribute software through any other system and a
1753
     licensee cannot impose that choice.
1754
 
1755
     This section is intended to make thoroughly clear what is believed
1756
     to be a consequence of the rest of this License.
1757
 
1758
  8. If the distribution and/or use of the Program is restricted in
1759
     certain countries either by patents or by copyrighted interfaces,
1760
     the original copyright holder who places the Program under this
1761
     License may add an explicit geographical distribution limitation
1762
     excluding those countries, so that distribution is permitted only
1763
     in or among countries not thus excluded.  In such case, this
1764
     License incorporates the limitation as if written in the body of
1765
     this License.
1766
 
1767
  9. The Free Software Foundation may publish revised and/or new
1768
     versions of the General Public License from time to time.  Such
1769
     new versions will be similar in spirit to the present version, but
1770
     may differ in detail to address new problems or concerns.
1771
 
1772
     Each version is given a distinguishing version number.  If the
1773
     Program specifies a version number of this License which applies
1774
     to it and "any later version", you have the option of following
1775
     the terms and conditions either of that version or of any later
1776
     version published by the Free Software Foundation.  If the Program
1777
     does not specify a version number of this License, you may choose
1778
     any version ever published by the Free Software Foundation.
1779
 
1780
 10. If you wish to incorporate parts of the Program into other free
1781
     programs whose distribution conditions are different, write to the
1782
     author to ask for permission.  For software which is copyrighted
1783
     by the Free Software Foundation, write to the Free Software
1784
     Foundation; we sometimes make exceptions for this.  Our decision
1785
     will be guided by the two goals of preserving the free status of
1786
     all derivatives of our free software and of promoting the sharing
1787
     and reuse of software generally.
1788
 
1789
                                NO WARRANTY
1790
 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO
1791
     WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE
1792
     LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
1793
     HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT
1794
     WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT
1795
     NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
1796
     FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS TO THE
1797
     QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
1798
     PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY
1799
     SERVICING, REPAIR OR CORRECTION.
1800
 
1801
 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
1802
     WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY
1803
     MODIFY AND/OR REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE
1804
     LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL,
1805
     INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR
1806
     INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
1807
     DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU
1808
     OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY
1809
     OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN
1810
     ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
1811
 
1812
                      END OF TERMS AND CONDITIONS
1813
Appendix: How to Apply These Terms to Your New Programs
1814
=======================================================
1815
 
1816
If you develop a new program, and you want it to be of the greatest
1817
possible use to the public, the best way to achieve this is to make it
1818
free software which everyone can redistribute and change under these
1819
terms.
1820
 
1821
   To do so, attach the following notices to the program.  It is safest
1822
to attach them to the start of each source file to most effectively
1823
convey the exclusion of warranty; and each file should have at least
1824
the "copyright" line and a pointer to where the full notice is found.
1825
 
1826
     ONE LINE TO GIVE THE PROGRAM'S NAME AND A BRIEF IDEA OF WHAT IT DOES.
1827
     Copyright (C) YEAR  NAME OF AUTHOR
1828
 
1829
     This program is free software; you can redistribute it and/or modify
1830
     it under the terms of the GNU General Public License as published by
1831
     the Free Software Foundation; either version 2 of the License, or
1832
     (at your option) any later version.
1833
 
1834
     This program is distributed in the hope that it will be useful,
1835
     but WITHOUT ANY WARRANTY; without even the implied warranty of
1836
     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
1837
     GNU General Public License for more details.
1838
 
1839
     You should have received a copy of the GNU General Public License
1840
     along with this program; if not, write to the Free Software
1841
     Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
1842
 
1843
   Also add information on how to contact you by electronic and paper
1844
mail.
1845
 
1846
   If the program is interactive, make it output a short notice like
1847
this when it starts in an interactive mode:
1848
 
1849
     Gnomovision version 69, Copyright (C) YEAR NAME OF AUTHOR
1850
     Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
1851
     type `show w'.
1852
     This is free software, and you are welcome to redistribute it
1853
     under certain conditions; type `show c' for details.
1854
 
1855
   The hypothetical commands `show w' and `show c' should show the
1856
appropriate parts of the General Public License.  Of course, the
1857
commands you use may be called something other than `show w' and `show
1858
c'; they could even be mouse-clicks or menu items--whatever suits your
1859
program.
1860
 
1861
   You should also get your employer (if you work as a programmer) or
1862
your school, if any, to sign a "copyright disclaimer" for the program,
1863
if necessary.  Here is a sample; alter the names:
1864
 
1865
     Yoyodyne, Inc., hereby disclaims all copyright interest in the program
1866
     `Gnomovision' (which makes passes at compilers) written by James Hacker.
1867
 
1868
     SIGNATURE OF TY COON, 1 April 1989
1869
     Ty Coon, President of Vice
1870
 
1871
   This General Public License does not permit incorporating your
1872
program into proprietary programs.  If your program is a subroutine
1873
library, you may consider it more useful to permit linking proprietary
1874
applications with the library.  If this is what you want to do, use the
1875
GNU Library General Public License instead of this License.
1876
 
1877

1878
File: libgomp.info,  Node: GNU Free Documentation License,  Next: Funding,  Prev: Copying,  Up: Top
1879
 
1880
GNU Free Documentation License
1881
******************************
1882
 
1883
                      Version 1.2, November 2002
1884
 
1885
     Copyright (C) 2000,2001,2002 Free Software Foundation, Inc.
1886
     51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
1887
 
1888
     Everyone is permitted to copy and distribute verbatim copies
1889
     of this license document, but changing it is not allowed.
1890
 
1891
  0. PREAMBLE
1892
 
1893
     The purpose of this License is to make a manual, textbook, or other
1894
     functional and useful document "free" in the sense of freedom: to
1895
     assure everyone the effective freedom to copy and redistribute it,
1896
     with or without modifying it, either commercially or
1897
     noncommercially.  Secondarily, this License preserves for the
1898
     author and publisher a way to get credit for their work, while not
1899
     being considered responsible for modifications made by others.
1900
 
1901
     This License is a kind of "copyleft", which means that derivative
1902
     works of the document must themselves be free in the same sense.
1903
     It complements the GNU General Public License, which is a copyleft
1904
     license designed for free software.
1905
 
1906
     We have designed this License in order to use it for manuals for
1907
     free software, because free software needs free documentation: a
1908
     free program should come with manuals providing the same freedoms
1909
     that the software does.  But this License is not limited to
1910
     software manuals; it can be used for any textual work, regardless
1911
     of subject matter or whether it is published as a printed book.
1912
     We recommend this License principally for works whose purpose is
1913
     instruction or reference.
1914
 
1915
  1. APPLICABILITY AND DEFINITIONS
1916
 
1917
     This License applies to any manual or other work, in any medium,
1918
     that contains a notice placed by the copyright holder saying it
1919
     can be distributed under the terms of this License.  Such a notice
1920
     grants a world-wide, royalty-free license, unlimited in duration,
1921
     to use that work under the conditions stated herein.  The
1922
     "Document", below, refers to any such manual or work.  Any member
1923
     of the public is a licensee, and is addressed as "you".  You
1924
     accept the license if you copy, modify or distribute the work in a
1925
     way requiring permission under copyright law.
1926
 
1927
     A "Modified Version" of the Document means any work containing the
1928
     Document or a portion of it, either copied verbatim, or with
1929
     modifications and/or translated into another language.
1930
 
1931
     A "Secondary Section" is a named appendix or a front-matter section
1932
     of the Document that deals exclusively with the relationship of the
1933
     publishers or authors of the Document to the Document's overall
1934
     subject (or to related matters) and contains nothing that could
1935
     fall directly within that overall subject.  (Thus, if the Document
1936
     is in part a textbook of mathematics, a Secondary Section may not
1937
     explain any mathematics.)  The relationship could be a matter of
1938
     historical connection with the subject or with related matters, or
1939
     of legal, commercial, philosophical, ethical or political position
1940
     regarding them.
1941
 
1942
     The "Invariant Sections" are certain Secondary Sections whose
1943
     titles are designated, as being those of Invariant Sections, in
1944
     the notice that says that the Document is released under this
1945
     License.  If a section does not fit the above definition of
1946
     Secondary then it is not allowed to be designated as Invariant.
1947
     The Document may contain zero Invariant Sections.  If the Document
1948
     does not identify any Invariant Sections then there are none.
1949
 
1950
     The "Cover Texts" are certain short passages of text that are
1951
     listed, as Front-Cover Texts or Back-Cover Texts, in the notice
1952
     that says that the Document is released under this License.  A
1953
     Front-Cover Text may be at most 5 words, and a Back-Cover Text may
1954
     be at most 25 words.
1955
 
1956
     A "Transparent" copy of the Document means a machine-readable copy,
1957
     represented in a format whose specification is available to the
1958
     general public, that is suitable for revising the document
1959
     straightforwardly with generic text editors or (for images
1960
     composed of pixels) generic paint programs or (for drawings) some
1961
     widely available drawing editor, and that is suitable for input to
1962
     text formatters or for automatic translation to a variety of
1963
     formats suitable for input to text formatters.  A copy made in an
1964
     otherwise Transparent file format whose markup, or absence of
1965
     markup, has been arranged to thwart or discourage subsequent
1966
     modification by readers is not Transparent.  An image format is
1967
     not Transparent if used for any substantial amount of text.  A
1968
     copy that is not "Transparent" is called "Opaque".
1969
 
1970
     Examples of suitable formats for Transparent copies include plain
1971
     ASCII without markup, Texinfo input format, LaTeX input format,
1972
     SGML or XML using a publicly available DTD, and
1973
     standard-conforming simple HTML, PostScript or PDF designed for
1974
     human modification.  Examples of transparent image formats include
1975
     PNG, XCF and JPG.  Opaque formats include proprietary formats that
1976
     can be read and edited only by proprietary word processors, SGML or
1977
     XML for which the DTD and/or processing tools are not generally
1978
     available, and the machine-generated HTML, PostScript or PDF
1979
     produced by some word processors for output purposes only.
1980
 
1981
     The "Title Page" means, for a printed book, the title page itself,
1982
     plus such following pages as are needed to hold, legibly, the
1983
     material this License requires to appear in the title page.  For
1984
     works in formats which do not have any title page as such, "Title
1985
     Page" means the text near the most prominent appearance of the
1986
     work's title, preceding the beginning of the body of the text.
1987
 
1988
     A section "Entitled XYZ" means a named subunit of the Document
1989
     whose title either is precisely XYZ or contains XYZ in parentheses
1990
     following text that translates XYZ in another language.  (Here XYZ
1991
     stands for a specific section name mentioned below, such as
1992
     "Acknowledgements", "Dedications", "Endorsements", or "History".)
1993
     To "Preserve the Title" of such a section when you modify the
1994
     Document means that it remains a section "Entitled XYZ" according
1995
     to this definition.
1996
 
1997
     The Document may include Warranty Disclaimers next to the notice
1998
     which states that this License applies to the Document.  These
1999
     Warranty Disclaimers are considered to be included by reference in
2000
     this License, but only as regards disclaiming warranties: any other
2001
     implication that these Warranty Disclaimers may have is void and
2002
     has no effect on the meaning of this License.
2003
 
2004
  2. VERBATIM COPYING
2005
 
2006
     You may copy and distribute the Document in any medium, either
2007
     commercially or noncommercially, provided that this License, the
2008
     copyright notices, and the license notice saying this License
2009
     applies to the Document are reproduced in all copies, and that you
2010
     add no other conditions whatsoever to those of this License.  You
2011
     may not use technical measures to obstruct or control the reading
2012
     or further copying of the copies you make or distribute.  However,
2013
     you may accept compensation in exchange for copies.  If you
2014
     distribute a large enough number of copies you must also follow
2015
     the conditions in section 3.
2016
 
2017
     You may also lend copies, under the same conditions stated above,
2018
     and you may publicly display copies.
2019
 
2020
  3. COPYING IN QUANTITY
2021
 
2022
     If you publish printed copies (or copies in media that commonly
2023
     have printed covers) of the Document, numbering more than 100, and
2024
     the Document's license notice requires Cover Texts, you must
2025
     enclose the copies in covers that carry, clearly and legibly, all
2026
     these Cover Texts: Front-Cover Texts on the front cover, and
2027
     Back-Cover Texts on the back cover.  Both covers must also clearly
2028
     and legibly identify you as the publisher of these copies.  The
2029
     front cover must present the full title with all words of the
2030
     title equally prominent and visible.  You may add other material
2031
     on the covers in addition.  Copying with changes limited to the
2032
     covers, as long as they preserve the title of the Document and
2033
     satisfy these conditions, can be treated as verbatim copying in
2034
     other respects.
2035
 
2036
     If the required texts for either cover are too voluminous to fit
2037
     legibly, you should put the first ones listed (as many as fit
2038
     reasonably) on the actual cover, and continue the rest onto
2039
     adjacent pages.
2040
 
2041
     If you publish or distribute Opaque copies of the Document
2042
     numbering more than 100, you must either include a
2043
     machine-readable Transparent copy along with each Opaque copy, or
2044
     state in or with each Opaque copy a computer-network location from
2045
     which the general network-using public has access to download
2046
     using public-standard network protocols a complete Transparent
2047
     copy of the Document, free of added material.  If you use the
2048
     latter option, you must take reasonably prudent steps, when you
2049
     begin distribution of Opaque copies in quantity, to ensure that
2050
     this Transparent copy will remain thus accessible at the stated
2051
     location until at least one year after the last time you
2052
     distribute an Opaque copy (directly or through your agents or
2053
     retailers) of that edition to the public.
2054
 
2055
     It is requested, but not required, that you contact the authors of
2056
     the Document well before redistributing any large number of
2057
     copies, to give them a chance to provide you with an updated
2058
     version of the Document.
2059
 
2060
  4. MODIFICATIONS
2061
 
2062
     You may copy and distribute a Modified Version of the Document
2063
     under the conditions of sections 2 and 3 above, provided that you
2064
     release the Modified Version under precisely this License, with
2065
     the Modified Version filling the role of the Document, thus
2066
     licensing distribution and modification of the Modified Version to
2067
     whoever possesses a copy of it.  In addition, you must do these
2068
     things in the Modified Version:
2069
 
2070
       A. Use in the Title Page (and on the covers, if any) a title
2071
          distinct from that of the Document, and from those of
2072
          previous versions (which should, if there were any, be listed
2073
          in the History section of the Document).  You may use the
2074
          same title as a previous version if the original publisher of
2075
          that version gives permission.
2076
 
2077
       B. List on the Title Page, as authors, one or more persons or
2078
          entities responsible for authorship of the modifications in
2079
          the Modified Version, together with at least five of the
2080
          principal authors of the Document (all of its principal
2081
          authors, if it has fewer than five), unless they release you
2082
          from this requirement.
2083
 
2084
       C. State on the Title page the name of the publisher of the
2085
          Modified Version, as the publisher.
2086
 
2087
       D. Preserve all the copyright notices of the Document.
2088
 
2089
       E. Add an appropriate copyright notice for your modifications
2090
          adjacent to the other copyright notices.
2091
 
2092
       F. Include, immediately after the copyright notices, a license
2093
          notice giving the public permission to use the Modified
2094
          Version under the terms of this License, in the form shown in
2095
          the Addendum below.
2096
 
2097
       G. Preserve in that license notice the full lists of Invariant
2098
          Sections and required Cover Texts given in the Document's
2099
          license notice.
2100
 
2101
       H. Include an unaltered copy of this License.
2102
 
2103
       I. Preserve the section Entitled "History", Preserve its Title,
2104
          and add to it an item stating at least the title, year, new
2105
          authors, and publisher of the Modified Version as given on
2106
          the Title Page.  If there is no section Entitled "History" in
2107
          the Document, create one stating the title, year, authors,
2108
          and publisher of the Document as given on its Title Page,
2109
          then add an item describing the Modified Version as stated in
2110
          the previous sentence.
2111
 
2112
       J. Preserve the network location, if any, given in the Document
2113
          for public access to a Transparent copy of the Document, and
2114
          likewise the network locations given in the Document for
2115
          previous versions it was based on.  These may be placed in
2116
          the "History" section.  You may omit a network location for a
2117
          work that was published at least four years before the
2118
          Document itself, or if the original publisher of the version
2119
          it refers to gives permission.
2120
 
2121
       K. For any section Entitled "Acknowledgements" or "Dedications",
2122
          Preserve the Title of the section, and preserve in the
2123
          section all the substance and tone of each of the contributor
2124
          acknowledgements and/or dedications given therein.
2125
 
2126
       L. Preserve all the Invariant Sections of the Document,
2127
          unaltered in their text and in their titles.  Section numbers
2128
          or the equivalent are not considered part of the section
2129
          titles.
2130
 
2131
       M. Delete any section Entitled "Endorsements".  Such a section
2132
          may not be included in the Modified Version.
2133
 
2134
       N. Do not retitle any existing section to be Entitled
2135
          "Endorsements" or to conflict in title with any Invariant
2136
          Section.
2137
 
2138
       O. Preserve any Warranty Disclaimers.
2139
 
2140
     If the Modified Version includes new front-matter sections or
2141
     appendices that qualify as Secondary Sections and contain no
2142
     material copied from the Document, you may at your option
2143
     designate some or all of these sections as invariant.  To do this,
2144
     add their titles to the list of Invariant Sections in the Modified
2145
     Version's license notice.  These titles must be distinct from any
2146
     other section titles.
2147
 
2148
     You may add a section Entitled "Endorsements", provided it contains
2149
     nothing but endorsements of your Modified Version by various
2150
     parties--for example, statements of peer review or that the text
2151
     has been approved by an organization as the authoritative
2152
     definition of a standard.
2153
 
2154
     You may add a passage of up to five words as a Front-Cover Text,
2155
     and a passage of up to 25 words as a Back-Cover Text, to the end
2156
     of the list of Cover Texts in the Modified Version.  Only one
2157
     passage of Front-Cover Text and one of Back-Cover Text may be
2158
     added by (or through arrangements made by) any one entity.  If the
2159
     Document already includes a cover text for the same cover,
2160
     previously added by you or by arrangement made by the same entity
2161
     you are acting on behalf of, you may not add another; but you may
2162
     replace the old one, on explicit permission from the previous
2163
     publisher that added the old one.
2164
 
2165
     The author(s) and publisher(s) of the Document do not by this
2166
     License give permission to use their names for publicity for or to
2167
     assert or imply endorsement of any Modified Version.
2168
 
2169
  5. COMBINING DOCUMENTS
2170
 
2171
     You may combine the Document with other documents released under
2172
     this License, under the terms defined in section 4 above for
2173
     modified versions, provided that you include in the combination
2174
     all of the Invariant Sections of all of the original documents,
2175
     unmodified, and list them all as Invariant Sections of your
2176
     combined work in its license notice, and that you preserve all
2177
     their Warranty Disclaimers.
2178
 
2179
     The combined work need only contain one copy of this License, and
2180
     multiple identical Invariant Sections may be replaced with a single
2181
     copy.  If there are multiple Invariant Sections with the same name
2182
     but different contents, make the title of each such section unique
2183
     by adding at the end of it, in parentheses, the name of the
2184
     original author or publisher of that section if known, or else a
2185
     unique number.  Make the same adjustment to the section titles in
2186
     the list of Invariant Sections in the license notice of the
2187
     combined work.
2188
 
2189
     In the combination, you must combine any sections Entitled
2190
     "History" in the various original documents, forming one section
2191
     Entitled "History"; likewise combine any sections Entitled
2192
     "Acknowledgements", and any sections Entitled "Dedications".  You
2193
     must delete all sections Entitled "Endorsements."
2194
 
2195
  6. COLLECTIONS OF DOCUMENTS
2196
 
2197
     You may make a collection consisting of the Document and other
2198
     documents released under this License, and replace the individual
2199
     copies of this License in the various documents with a single copy
2200
     that is included in the collection, provided that you follow the
2201
     rules of this License for verbatim copying of each of the
2202
     documents in all other respects.
2203
 
2204
     You may extract a single document from such a collection, and
2205
     distribute it individually under this License, provided you insert
2206
     a copy of this License into the extracted document, and follow
2207
     this License in all other respects regarding verbatim copying of
2208
     that document.
2209
 
2210
  7. AGGREGATION WITH INDEPENDENT WORKS
2211
 
2212
     A compilation of the Document or its derivatives with other
2213
     separate and independent documents or works, in or on a volume of
2214
     a storage or distribution medium, is called an "aggregate" if the
2215
     copyright resulting from the compilation is not used to limit the
2216
     legal rights of the compilation's users beyond what the individual
2217
     works permit.  When the Document is included in an aggregate, this
2218
     License does not apply to the other works in the aggregate which
2219
     are not themselves derivative works of the Document.
2220
 
2221
     If the Cover Text requirement of section 3 is applicable to these
2222
     copies of the Document, then if the Document is less than one half
2223
     of the entire aggregate, the Document's Cover Texts may be placed
2224
     on covers that bracket the Document within the aggregate, or the
2225
     electronic equivalent of covers if the Document is in electronic
2226
     form.  Otherwise they must appear on printed covers that bracket
2227
     the whole aggregate.
2228
 
2229
  8. TRANSLATION
2230
 
2231
     Translation is considered a kind of modification, so you may
2232
     distribute translations of the Document under the terms of section
2233
     4.  Replacing Invariant Sections with translations requires special
2234
     permission from their copyright holders, but you may include
2235
     translations of some or all Invariant Sections in addition to the
2236
     original versions of these Invariant Sections.  You may include a
2237
     translation of this License, and all the license notices in the
2238
     Document, and any Warranty Disclaimers, provided that you also
2239
     include the original English version of this License and the
2240
     original versions of those notices and disclaimers.  In case of a
2241
     disagreement between the translation and the original version of
2242
     this License or a notice or disclaimer, the original version will
2243
     prevail.
2244
 
2245
     If a section in the Document is Entitled "Acknowledgements",
2246
     "Dedications", or "History", the requirement (section 4) to
2247
     Preserve its Title (section 1) will typically require changing the
2248
     actual title.
2249
 
2250
  9. TERMINATION
2251
 
2252
     You may not copy, modify, sublicense, or distribute the Document
2253
     except as expressly provided for under this License.  Any other
2254
     attempt to copy, modify, sublicense or distribute the Document is
2255
     void, and will automatically terminate your rights under this
2256
     License.  However, parties who have received copies, or rights,
2257
     from you under this License will not have their licenses
2258
     terminated so long as such parties remain in full compliance.
2259
 
2260
 10. FUTURE REVISIONS OF THIS LICENSE
2261
 
2262
     The Free Software Foundation may publish new, revised versions of
2263
     the GNU Free Documentation License from time to time.  Such new
2264
     versions will be similar in spirit to the present version, but may
2265
     differ in detail to address new problems or concerns.  See
2266
     `http://www.gnu.org/copyleft/'.
2267
 
2268
     Each version of the License is given a distinguishing version
2269
     number.  If the Document specifies that a particular numbered
2270
     version of this License "or any later version" applies to it, you
2271
     have the option of following the terms and conditions either of
2272
     that specified version or of any later version that has been
2273
     published (not as a draft) by the Free Software Foundation.  If
2274
     the Document does not specify a version number of this License,
2275
     you may choose any version ever published (not as a draft) by the
2276
     Free Software Foundation.
2277
 
2278
ADDENDUM: How to use this License for your documents
2279
====================================================
2280
 
2281
To use this License in a document you have written, include a copy of
2282
the License in the document and put the following copyright and license
2283
notices just after the title page:
2284
 
2285
       Copyright (C)  YEAR  YOUR NAME.
2286
       Permission is granted to copy, distribute and/or modify this document
2287
       under the terms of the GNU Free Documentation License, Version 1.2
2288
       or any later version published by the Free Software Foundation;
2289
       with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
2290
       Texts.  A copy of the license is included in the section entitled ``GNU
2291
       Free Documentation License''.
2292
 
2293
   If you have Invariant Sections, Front-Cover Texts and Back-Cover
2294
Texts, replace the "with...Texts." line with this:
2295
 
2296
         with the Invariant Sections being LIST THEIR TITLES, with
2297
         the Front-Cover Texts being LIST, and with the Back-Cover Texts
2298
         being LIST.
2299
 
2300
   If you have Invariant Sections without Cover Texts, or some other
2301
combination of the three, merge those two alternatives to suit the
2302
situation.
2303
 
2304
   If your document contains nontrivial examples of program code, we
2305
recommend releasing these examples in parallel under your choice of
2306
free software license, such as the GNU General Public License, to
2307
permit their use in free software.
2308
 
2309

2310
File: libgomp.info,  Node: Funding,  Next: Index,  Prev: GNU Free Documentation License,  Up: Top
2311
 
2312
Funding Free Software
2313
*********************
2314
 
2315
If you want to have more free software a few years from now, it makes
2316
sense for you to help encourage people to contribute funds for its
2317
development.  The most effective approach known is to encourage
2318
commercial redistributors to donate.
2319
 
2320
   Users of free software systems can boost the pace of development by
2321
encouraging for-a-fee distributors to donate part of their selling price
2322
to free software developers--the Free Software Foundation, and others.
2323
 
2324
   The way to convince distributors to do this is to demand it and
2325
expect it from them.  So when you compare distributors, judge them
2326
partly by how much they give to free software development.  Show
2327
distributors they must compete to be the one who gives the most.
2328
 
2329
   To make this approach work, you must insist on numbers that you can
2330
compare, such as, "We will donate ten dollars to the Frobnitz project
2331
for each disk sold."  Don't be satisfied with a vague promise, such as
2332
"A portion of the profits are donated," since it doesn't give a basis
2333
for comparison.
2334
 
2335
   Even a precise fraction "of the profits from this disk" is not very
2336
meaningful, since creative accounting and unrelated business decisions
2337
can greatly alter what fraction of the sales price counts as profit.
2338
If the price you pay is $50, ten percent of the profit is probably less
2339
than a dollar; it might be a few cents, or nothing at all.
2340
 
2341
   Some redistributors do development work themselves.  This is useful
2342
too; but to keep everyone honest, you need to inquire how much they do,
2343
and what kind.  Some kinds of development make much more long-term
2344
difference than others.  For example, maintaining a separate version of
2345
a program contributes very little; maintaining the standard version of a
2346
program for the whole community contributes much.  Easy new ports
2347
contribute little, since someone else would surely do them; difficult
2348
ports such as adding a new CPU to the GNU Compiler Collection
2349
contribute more; major new features or packages contribute the most.
2350
 
2351
   By establishing the idea that supporting further development is "the
2352
proper thing to do" when distributing free software for a fee, we can
2353
assure a steady flow of resources into making more free software.
2354
 
2355
     Copyright (C) 1994 Free Software Foundation, Inc.
2356
     Verbatim copying and redistribution of this section is permitted
2357
     without royalty; alteration is not permitted.
2358
 
2359

2360
File: libgomp.info,  Node: Index,  Prev: Funding,  Up: Top
2361
 
2362
Index
2363
*****
2364
 
2365
 
2366
* Menu:
2367
2368
* Environment Variable <1>:              GOMP_STACKSIZE.        (line 6)
2369
* Environment Variable <2>:              GOMP_CPU_AFFINITY.     (line 6)
2370
* Environment Variable <3>:              OMP_WAIT_POLICY.       (line 6)
2371
* Environment Variable <4>:              OMP_THREAD_LIMIT.      (line 6)
2372
* Environment Variable <5>:              OMP_STACKSIZE.         (line 6)
2373
* Environment Variable <6>:              OMP_SCHEDULE.          (line 6)
2374
* Environment Variable <7>:              OMP_NUM_THREADS.       (line 6)
2375
* Environment Variable <8>:              OMP_NESTED.            (line 6)
2376
* Environment Variable <9>:              OMP_MAX_ACTIVE_LEVELS. (line 6)
2377
* Environment Variable:                  OMP_DYNAMIC.           (line 6)
2378
* FDL, GNU Free Documentation License:   GNU Free Documentation License.
2379
                                                                (line 6)
2380
* Implementation specific setting <1>:   GOMP_STACKSIZE.        (line 6)
2381
* Implementation specific setting <2>:   OMP_SCHEDULE.          (line 6)
2382
 
2383
 
2384
* Introduction:                          Top.                   (line 6)
2385
2386
2387

2388
Tag Table:
2389
Node: Top2053
2390
Node: Enabling OpenMP3247
2391
Node: Runtime Library Routines4032
2392
Node: omp_get_active_level6407
2393
Node: omp_get_ancestor_thread_num7098
2394
Node: omp_get_dynamic8012
2395
Node: omp_get_level8886
2396
Node: omp_get_max_active_levels9497
2397
Node: omp_get_max_threads10185
2398
Node: omp_get_nested10937
2399
Node: omp_get_num_procs11845
2400
Node: omp_get_num_threads12359
2401
Node: omp_get_schedule13429
2402
Node: omp_get_team_size14336
2403
Node: omp_get_thread_limit15294
2404
Node: omp_get_thread_num15913
2405
Node: omp_in_parallel16767
2406
Node: omp_set_dynamic17413
2407
Node: omp_set_max_active_levels18249
2408
Node: omp_set_nested19011
2409
Node: omp_set_num_threads19888
2410
Node: omp_set_schedule20726
2411
Node: omp_init_lock21770
2412
Node: omp_set_lock22420
2413
Node: omp_test_lock23269
2414
Node: omp_unset_lock24296
2415
Node: omp_destroy_lock25222
2416
Node: omp_init_nest_lock25892
2417
Node: omp_set_nest_lock26624
2418
Node: omp_test_nest_lock27533
2419
Node: omp_unset_nest_lock28631
2420
Node: omp_destroy_nest_lock29640
2421
Node: omp_get_wtick30388
2422
Node: omp_get_wtime30975
2423
Node: Environment Variables31758
2424
Node: OMP_DYNAMIC32819
2425
Node: OMP_MAX_ACTIVE_LEVELS33387
2426
Node: OMP_NESTED34024
2427
Node: OMP_NUM_THREADS34628
2428
Node: OMP_SCHEDULE35201
2429
Node: OMP_STACKSIZE35895
2430
Node: OMP_THREAD_LIMIT36720
2431
Node: OMP_WAIT_POLICY37313
2432
Node: GOMP_CPU_AFFINITY37878
2433
Node: GOMP_STACKSIZE39362
2434
Node: The libgomp ABI40172
2435
Node: Implementing MASTER construct40970
2436
Node: Implementing CRITICAL construct41383
2437
Node: Implementing ATOMIC construct42131
2438
Node: Implementing FLUSH construct42612
2439
Node: Implementing BARRIER construct42883
2440
Node: Implementing THREADPRIVATE construct43152
2441
Node: Implementing PRIVATE clause43804
2442
Node: Implementing FIRSTPRIVATE LASTPRIVATE COPYIN and COPYPRIVATE clauses44385
2443
Node: Implementing REDUCTION clause45700
2444
Node: Implementing PARALLEL construct46256
2445
Node: Implementing FOR construct47513
2446
Node: Implementing ORDERED construct49511
2447
Node: Implementing SECTIONS construct49817
2448
Node: Implementing SINGLE construct50583
2449
Node: Reporting Bugs51245
2450
Node: Copying51553
2451
Node: GNU Free Documentation License70763
2452
Node: Funding93174

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.