gcc/libgomp/loop.c

/* Copyright (C) 2005, 2008 Free Software Foundation, Inc.
   Contributed by Richard Henderson <rth@redhat.com>.

   This file is part of the GNU OpenMP Library (libgomp).

   Libgomp is free software; you can redistribute it and/or modify it
   under the terms of the GNU Lesser General Public License as published by
   the Free Software Foundation; either version 2.1 of the License, or
   (at your option) any later version.

   Libgomp is distributed in the hope that it will be useful, but WITHOUT ANY
   WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
   FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public License for
   more details.

   You should have received a copy of the GNU Lesser General Public License 
   along with libgomp; see the file COPYING.LIB.  If not, write to the
   Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
   MA 02110-1301, USA.  */

/* As a special exception, if you link this library with other files, some
   of which are compiled with GCC, to produce an executable, this library
   does not by itself cause the resulting executable to be covered by the
   GNU General Public License.  This exception does not however invalidate
   any other reasons why the executable file might be covered by the GNU
   General Public License.  */

/* This file handles the LOOP (FOR/DO) construct.  */

#include <limits.h>
#include <stdlib.h>
#include "libgomp.h"


/* Initialize the given work share construct from the given arguments.  */

static inline void
gomp_loop_init (struct gomp_work_share *ws, long start, long end, long incr,
		enum gomp_schedule_type sched, long chunk_size)
{
  ws->sched = sched;
  ws->chunk_size = chunk_size;
  /* Canonicalize loops that have zero iterations to ->next == ->end.  */
  ws->end = ((incr > 0 && start > end) || (incr < 0 && start < end))
	    ? start : end;
  ws->incr = incr;
  ws->next = start;
  if (sched == GFS_DYNAMIC)
    {
      ws->chunk_size *= incr;

#ifdef HAVE_SYNC_BUILTINS
      {
	/* For dynamic scheduling prepare things to make each iteration
	   faster.  */
	struct gomp_thread *thr = gomp_thread ();
	struct gomp_team *team = thr->ts.team;
	long nthreads = team ? team->nthreads : 1;

	if (__builtin_expect (incr > 0, 1))
	  {
	    /* Cheap overflow protection.  */
	    if (__builtin_expect ((nthreads | ws->chunk_size)
				  >= 1UL << (sizeof (long)
					     * __CHAR_BIT__ / 2 - 1), 0))
	      ws->mode = 0;
	    else
	      ws->mode = ws->end < (LONG_MAX
				    - (nthreads + 1) * ws->chunk_size);
	  }
	/* Cheap overflow protection.  */
	else if (__builtin_expect ((nthreads | -ws->chunk_size)
				   >= 1UL << (sizeof (long)
					      * __CHAR_BIT__ / 2 - 1), 0))
	  ws->mode = 0;
	else
	  ws->mode = ws->end > (nthreads + 1) * -ws->chunk_size - LONG_MAX;
      }
#endif
    }
}

/* The *_start routines are called when first encountering a loop construct
   that is not bound directly to a parallel construct.  The first thread 
   that arrives will create the work-share construct; subsequent threads
   will see the construct exists and allocate work from it.

   START, END, INCR are the bounds of the loop; due to the restrictions of
   OpenMP, these values must be the same in every thread.  This is not 
   verified (nor is it entirely verifiable, since START is not necessarily
   retained intact in the work-share data structure).  CHUNK_SIZE is the
   scheduling parameter; again this must be identical in all threads.

   Returns true if there's any work for this thread to perform.  If so,
   *ISTART and *IEND are filled with the bounds of the iteration block
   allocated to this thread.  Returns false if all work was assigned to
   other threads prior to this thread's arrival.  */

static bool
gomp_loop_static_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (gomp_work_share_start (false))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_STATIC, chunk_size);
      gomp_work_share_init_done ();
    }

  return !gomp_iter_static_next (istart, iend);
}

static bool
gomp_loop_dynamic_start (long start, long end, long incr, long chunk_size,
			 long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (false))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_DYNAMIC, chunk_size);
      gomp_work_share_init_done ();
    }

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_dynamic_next (istart, iend);
#else
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

static bool
gomp_loop_guided_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (false))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_GUIDED, chunk_size);
      gomp_work_share_init_done ();
    }

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_guided_next (istart, iend);
#else
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

bool
GOMP_loop_runtime_start (long start, long end, long incr,
			 long *istart, long *iend)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  switch (icv->run_sched_var)
    {
    case GFS_STATIC:
      return gomp_loop_static_start (start, end, incr, icv->run_sched_modifier,
				     istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_dynamic_start (start, end, incr, icv->run_sched_modifier,
				      istart, iend);
    case GFS_GUIDED:
      return gomp_loop_guided_start (start, end, incr, icv->run_sched_modifier,
				     istart, iend);
    case GFS_AUTO:
      /* For now map to schedule(static), later on we could play with feedback
	 driven choice.  */
      return gomp_loop_static_start (start, end, incr, 0, istart, iend);
    default:
      abort ();
    }
}

/* The *_ordered_*_start routines are similar.  The only difference is that
   this work-share construct is initialized to expect an ORDERED section.  */

static bool
gomp_loop_ordered_static_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (gomp_work_share_start (true))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_STATIC, chunk_size);
      gomp_ordered_static_init ();
      gomp_work_share_init_done ();
    }

  return !gomp_iter_static_next (istart, iend);
}

static bool
gomp_loop_ordered_dynamic_start (long start, long end, long incr,
				 long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (true))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_DYNAMIC, chunk_size);
      gomp_mutex_lock (&thr->ts.work_share->lock);
      gomp_work_share_init_done ();
    }
  else
    gomp_mutex_lock (&thr->ts.work_share->lock);

  ret = gomp_iter_dynamic_next_locked (istart, iend);
  if (ret)
    gomp_ordered_first ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

static bool
gomp_loop_ordered_guided_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (true))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_GUIDED, chunk_size);
      gomp_mutex_lock (&thr->ts.work_share->lock);
      gomp_work_share_init_done ();
    }
  else
    gomp_mutex_lock (&thr->ts.work_share->lock);

  ret = gomp_iter_guided_next_locked (istart, iend);
  if (ret)
    gomp_ordered_first ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

bool
GOMP_loop_ordered_runtime_start (long start, long end, long incr,
				 long *istart, long *iend)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  switch (icv->run_sched_var)
    {
    case GFS_STATIC:
      return gomp_loop_ordered_static_start (start, end, incr,
					     icv->run_sched_modifier,
					     istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_ordered_dynamic_start (start, end, incr,
					      icv->run_sched_modifier,
					      istart, iend);
    case GFS_GUIDED:
      return gomp_loop_ordered_guided_start (start, end, incr,
					     icv->run_sched_modifier,
					     istart, iend);
    case GFS_AUTO:
      /* For now map to schedule(static), later on we could play with feedback
	 driven choice.  */
      return gomp_loop_ordered_static_start (start, end, incr,
					     0, istart, iend);
    default:
      abort ();
    }
}

/* The *_next routines are called when the thread completes processing of 
   the iteration block currently assigned to it.  If the work-share 
   construct is bound directly to a parallel construct, then the iteration
   bounds may have been set up before the parallel.  In which case, this
   may be the first iteration for the thread.

   Returns true if there is work remaining to be performed; *ISTART and
   *IEND are filled with a new iteration block.  Returns false if all work
   has been assigned.  */

static bool
gomp_loop_static_next (long *istart, long *iend)
{
  return !gomp_iter_static_next (istart, iend);
}

static bool
gomp_loop_dynamic_next (long *istart, long *iend)
{
  bool ret;

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_dynamic_next (istart, iend);
#else
  struct gomp_thread *thr = gomp_thread ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

static bool
gomp_loop_guided_next (long *istart, long *iend)
{
  bool ret;

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_guided_next (istart, iend);
#else
  struct gomp_thread *thr = gomp_thread ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

bool
GOMP_loop_runtime_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  
  switch (thr->ts.work_share->sched)
    {
    case GFS_STATIC:
    case GFS_AUTO:
      return gomp_loop_static_next (istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_dynamic_next (istart, iend);
    case GFS_GUIDED:
      return gomp_loop_guided_next (istart, iend);
    default:
      abort ();
    }
}

/* The *_ordered_*_next routines are called when the thread completes
   processing of the iteration block currently assigned to it.

   Returns true if there is work remaining to be performed; *ISTART and
   *IEND are filled with a new iteration block.  Returns false if all work
   has been assigned.  */

static bool
gomp_loop_ordered_static_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  int test;

  gomp_ordered_sync ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  test = gomp_iter_static_next (istart, iend);
  if (test >= 0)
    gomp_ordered_static_next ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return test == 0;
}

static bool
gomp_loop_ordered_dynamic_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  gomp_ordered_sync ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  if (ret)
    gomp_ordered_next ();
  else
    gomp_ordered_last ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

static bool
gomp_loop_ordered_guided_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  gomp_ordered_sync ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  if (ret)
    gomp_ordered_next ();
  else
    gomp_ordered_last ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

bool
GOMP_loop_ordered_runtime_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  
  switch (thr->ts.work_share->sched)
    {
    case GFS_STATIC:
    case GFS_AUTO:
      return gomp_loop_ordered_static_next (istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_ordered_dynamic_next (istart, iend);
    case GFS_GUIDED:
      return gomp_loop_ordered_guided_next (istart, iend);
    default:
      abort ();
    }
}

/* The GOMP_parallel_loop_* routines pre-initialize a work-share construct
   to avoid one synchronization once we get into the loop.  */

static void
gomp_parallel_loop_start (void (*fn) (void *), void *data,
			  unsigned num_threads, long start, long end,
			  long incr, enum gomp_schedule_type sched,
			  long chunk_size)
{
  struct gomp_team *team;

  num_threads = gomp_resolve_num_threads (num_threads, 0);
  team = gomp_new_team (num_threads);
  gomp_loop_init (&team->work_shares[0], start, end, incr, sched, chunk_size);
  gomp_team_start (fn, data, num_threads, team);
}

void
GOMP_parallel_loop_static_start (void (*fn) (void *), void *data,
				 unsigned num_threads, long start, long end,
				 long incr, long chunk_size)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_STATIC, chunk_size);
}

void
GOMP_parallel_loop_dynamic_start (void (*fn) (void *), void *data,
				  unsigned num_threads, long start, long end,
				  long incr, long chunk_size)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_DYNAMIC, chunk_size);
}

void
GOMP_parallel_loop_guided_start (void (*fn) (void *), void *data,
				 unsigned num_threads, long start, long end,
				 long incr, long chunk_size)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_GUIDED, chunk_size);
}

void
GOMP_parallel_loop_runtime_start (void (*fn) (void *), void *data,
				  unsigned num_threads, long start, long end,
				  long incr)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    icv->run_sched_var, icv->run_sched_modifier);
}

/* The GOMP_loop_end* routines are called after the thread is told that
   all loop iterations are complete.  This first version synchronizes
   all threads; the nowait version does not.  */

void
GOMP_loop_end (void)
{
  gomp_work_share_end ();
}

void
GOMP_loop_end_nowait (void)
{
  gomp_work_share_end_nowait ();
}


/* We use static functions above so that we're sure that the "runtime"
   function can defer to the proper routine without interposition.  We
   export the static function with a strong alias when possible, or with
   a wrapper function otherwise.  */

#ifdef HAVE_ATTRIBUTE_ALIAS
extern __typeof(gomp_loop_static_start) GOMP_loop_static_start
	__attribute__((alias ("gomp_loop_static_start")));
extern __typeof(gomp_loop_dynamic_start) GOMP_loop_dynamic_start
	__attribute__((alias ("gomp_loop_dynamic_start")));
extern __typeof(gomp_loop_guided_start) GOMP_loop_guided_start
	__attribute__((alias ("gomp_loop_guided_start")));

extern __typeof(gomp_loop_ordered_static_start) GOMP_loop_ordered_static_start
	__attribute__((alias ("gomp_loop_ordered_static_start")));
extern __typeof(gomp_loop_ordered_dynamic_start) GOMP_loop_ordered_dynamic_start
	__attribute__((alias ("gomp_loop_ordered_dynamic_start")));
extern __typeof(gomp_loop_ordered_guided_start) GOMP_loop_ordered_guided_start
	__attribute__((alias ("gomp_loop_ordered_guided_start")));

extern __typeof(gomp_loop_static_next) GOMP_loop_static_next
	__attribute__((alias ("gomp_loop_static_next")));
extern __typeof(gomp_loop_dynamic_next) GOMP_loop_dynamic_next
	__attribute__((alias ("gomp_loop_dynamic_next")));
extern __typeof(gomp_loop_guided_next) GOMP_loop_guided_next
	__attribute__((alias ("gomp_loop_guided_next")));

extern __typeof(gomp_loop_ordered_static_next) GOMP_loop_ordered_static_next
	__attribute__((alias ("gomp_loop_ordered_static_next")));
extern __typeof(gomp_loop_ordered_dynamic_next) GOMP_loop_ordered_dynamic_next
	__attribute__((alias ("gomp_loop_ordered_dynamic_next")));
extern __typeof(gomp_loop_ordered_guided_next) GOMP_loop_ordered_guided_next
	__attribute__((alias ("gomp_loop_ordered_guided_next")));
#else
bool
GOMP_loop_static_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  return gomp_loop_static_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_dynamic_start (long start, long end, long incr, long chunk_size,
			 long *istart, long *iend)
{
  return gomp_loop_dynamic_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_guided_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  return gomp_loop_guided_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_ordered_static_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  return gomp_loop_ordered_static_start (start, end, incr, chunk_size,
					 istart, iend);
}

bool
GOMP_loop_ordered_dynamic_start (long start, long end, long incr,
				 long chunk_size, long *istart, long *iend)
{
  return gomp_loop_ordered_dynamic_start (start, end, incr, chunk_size,
					  istart, iend);
}

bool
GOMP_loop_ordered_guided_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  return gomp_loop_ordered_guided_start (start, end, incr, chunk_size,
					 istart, iend);
}

bool
GOMP_loop_static_next (long *istart, long *iend)
{
  return gomp_loop_static_next (istart, iend);
}

bool
GOMP_loop_dynamic_next (long *istart, long *iend)
{
  return gomp_loop_dynamic_next (istart, iend);
}

bool
GOMP_loop_guided_next (long *istart, long *iend)
{
  return gomp_loop_guided_next (istart, iend);
}

bool
GOMP_loop_ordered_static_next (long *istart, long *iend)
{
  return gomp_loop_ordered_static_next (istart, iend);
}

bool
GOMP_loop_ordered_dynamic_next (long *istart, long *iend)
{
  return gomp_loop_ordered_dynamic_next (istart, iend);
}

bool
GOMP_loop_ordered_guided_next (long *istart, long *iend)
{
  return gomp_loop_ordered_guided_next (istart, iend);
}
#endif