/* Copyright (C) 2005, 2006, 2007, 2008, 2009, 2010
|
/* Copyright (C) 2005, 2006, 2007, 2008, 2009, 2010
|
Free Software Foundation, Inc.
|
Free Software Foundation, Inc.
|
Contributed by Jakub Jelinek <jakub@redhat.com>.
|
Contributed by Jakub Jelinek <jakub@redhat.com>.
|
|
|
This file is part of the GNU OpenMP Library (libgomp).
|
This file is part of the GNU OpenMP Library (libgomp).
|
|
|
Libgomp is free software; you can redistribute it and/or modify it
|
Libgomp is free software; you can redistribute it and/or modify it
|
under the terms of the GNU General Public License as published by
|
under the terms of the GNU General Public License as published by
|
the Free Software Foundation; either version 3, or (at your option)
|
the Free Software Foundation; either version 3, or (at your option)
|
any later version.
|
any later version.
|
|
|
Libgomp is distributed in the hope that it will be useful, but WITHOUT ANY
|
Libgomp is distributed in the hope that it will be useful, but WITHOUT ANY
|
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
|
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
|
FOR A PARTICULAR PURPOSE. See the GNU General Public License for
|
FOR A PARTICULAR PURPOSE. See the GNU General Public License for
|
more details.
|
more details.
|
|
|
Under Section 7 of GPL version 3, you are granted additional
|
Under Section 7 of GPL version 3, you are granted additional
|
permissions described in the GCC Runtime Library Exception, version
|
permissions described in the GCC Runtime Library Exception, version
|
3.1, as published by the Free Software Foundation.
|
3.1, as published by the Free Software Foundation.
|
|
|
You should have received a copy of the GNU General Public License and
|
You should have received a copy of the GNU General Public License and
|
a copy of the GCC Runtime Library Exception along with this program;
|
a copy of the GCC Runtime Library Exception along with this program;
|
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
|
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
|
<http://www.gnu.org/licenses/>. */
|
<http://www.gnu.org/licenses/>. */
|
|
|
/* This file contains system specific routines related to counting
|
/* This file contains system specific routines related to counting
|
online processors and dynamic load balancing. */
|
online processors and dynamic load balancing. */
|
|
|
#ifndef _GNU_SOURCE
|
#ifndef _GNU_SOURCE
|
#define _GNU_SOURCE 1
|
#define _GNU_SOURCE 1
|
#endif
|
#endif
|
#include "libgomp.h"
|
#include "libgomp.h"
|
#include <sched.h>
|
#include <sched.h>
|
#include <stdlib.h>
|
#include <stdlib.h>
|
#include <unistd.h>
|
#include <unistd.h>
|
#ifdef HAVE_GETLOADAVG
|
#ifdef HAVE_GETLOADAVG
|
# ifdef HAVE_SYS_LOADAVG_H
|
# ifdef HAVE_SYS_LOADAVG_H
|
# include <sys/loadavg.h>
|
# include <sys/loadavg.h>
|
# endif
|
# endif
|
#endif
|
#endif
|
|
|
#ifdef HAVE_PTHREAD_AFFINITY_NP
|
#ifdef HAVE_PTHREAD_AFFINITY_NP
|
static unsigned long
|
static unsigned long
|
cpuset_popcount (cpu_set_t *cpusetp)
|
cpuset_popcount (cpu_set_t *cpusetp)
|
{
|
{
|
#ifdef CPU_COUNT
|
#ifdef CPU_COUNT
|
/* glibc 2.6 and above provide a macro for this. */
|
/* glibc 2.6 and above provide a macro for this. */
|
return CPU_COUNT (cpusetp);
|
return CPU_COUNT (cpusetp);
|
#else
|
#else
|
size_t i;
|
size_t i;
|
unsigned long ret = 0;
|
unsigned long ret = 0;
|
extern int check[sizeof (cpusetp->__bits[0]) == sizeof (unsigned long int)];
|
extern int check[sizeof (cpusetp->__bits[0]) == sizeof (unsigned long int)];
|
|
|
(void) check;
|
(void) check;
|
for (i = 0; i < sizeof (*cpusetp) / sizeof (cpusetp->__bits[0]); i++)
|
for (i = 0; i < sizeof (*cpusetp) / sizeof (cpusetp->__bits[0]); i++)
|
{
|
{
|
unsigned long int mask = cpusetp->__bits[i];
|
unsigned long int mask = cpusetp->__bits[i];
|
if (mask == 0)
|
if (mask == 0)
|
continue;
|
continue;
|
ret += __builtin_popcountl (mask);
|
ret += __builtin_popcountl (mask);
|
}
|
}
|
return ret;
|
return ret;
|
#endif
|
#endif
|
}
|
}
|
#endif
|
#endif
|
|
|
/* At startup, determine the default number of threads. It would seem
|
/* At startup, determine the default number of threads. It would seem
|
this should be related to the number of cpus online. */
|
this should be related to the number of cpus online. */
|
|
|
void
|
void
|
gomp_init_num_threads (void)
|
gomp_init_num_threads (void)
|
{
|
{
|
#ifdef HAVE_PTHREAD_AFFINITY_NP
|
#ifdef HAVE_PTHREAD_AFFINITY_NP
|
cpu_set_t cpuset;
|
cpu_set_t cpuset;
|
|
|
if (pthread_getaffinity_np (pthread_self (), sizeof (cpuset), &cpuset) == 0)
|
if (pthread_getaffinity_np (pthread_self (), sizeof (cpuset), &cpuset) == 0)
|
{
|
{
|
/* Count only the CPUs this process can use. */
|
/* Count only the CPUs this process can use. */
|
gomp_global_icv.nthreads_var = cpuset_popcount (&cpuset);
|
gomp_global_icv.nthreads_var = cpuset_popcount (&cpuset);
|
if (gomp_global_icv.nthreads_var == 0)
|
if (gomp_global_icv.nthreads_var == 0)
|
gomp_global_icv.nthreads_var = 1;
|
gomp_global_icv.nthreads_var = 1;
|
return;
|
return;
|
}
|
}
|
#endif
|
#endif
|
#ifdef _SC_NPROCESSORS_ONLN
|
#ifdef _SC_NPROCESSORS_ONLN
|
gomp_global_icv.nthreads_var = sysconf (_SC_NPROCESSORS_ONLN);
|
gomp_global_icv.nthreads_var = sysconf (_SC_NPROCESSORS_ONLN);
|
#endif
|
#endif
|
}
|
}
|
|
|
static int
|
static int
|
get_num_procs (void)
|
get_num_procs (void)
|
{
|
{
|
#ifdef HAVE_PTHREAD_AFFINITY_NP
|
#ifdef HAVE_PTHREAD_AFFINITY_NP
|
cpu_set_t cpuset;
|
cpu_set_t cpuset;
|
|
|
if (gomp_cpu_affinity == NULL)
|
if (gomp_cpu_affinity == NULL)
|
{
|
{
|
/* Count only the CPUs this process can use. */
|
/* Count only the CPUs this process can use. */
|
if (pthread_getaffinity_np (pthread_self (), sizeof (cpuset),
|
if (pthread_getaffinity_np (pthread_self (), sizeof (cpuset),
|
&cpuset) == 0)
|
&cpuset) == 0)
|
{
|
{
|
int ret = cpuset_popcount (&cpuset);
|
int ret = cpuset_popcount (&cpuset);
|
return ret != 0 ? ret : 1;
|
return ret != 0 ? ret : 1;
|
}
|
}
|
}
|
}
|
else
|
else
|
{
|
{
|
/* We can't use pthread_getaffinity_np in this case
|
/* We can't use pthread_getaffinity_np in this case
|
(we have changed it ourselves, it binds to just one CPU).
|
(we have changed it ourselves, it binds to just one CPU).
|
Count instead the number of different CPUs we are
|
Count instead the number of different CPUs we are
|
using. gomp_init_affinity updated gomp_available_cpus to
|
using. gomp_init_affinity updated gomp_available_cpus to
|
the number of CPUs in the GOMP_AFFINITY mask that we are
|
the number of CPUs in the GOMP_AFFINITY mask that we are
|
allowed to use though. */
|
allowed to use though. */
|
return gomp_available_cpus;
|
return gomp_available_cpus;
|
}
|
}
|
#endif
|
#endif
|
#ifdef _SC_NPROCESSORS_ONLN
|
#ifdef _SC_NPROCESSORS_ONLN
|
return sysconf (_SC_NPROCESSORS_ONLN);
|
return sysconf (_SC_NPROCESSORS_ONLN);
|
#else
|
#else
|
return gomp_icv (false)->nthreads_var;
|
return gomp_icv (false)->nthreads_var;
|
#endif
|
#endif
|
}
|
}
|
|
|
/* When OMP_DYNAMIC is set, at thread launch determine the number of
|
/* When OMP_DYNAMIC is set, at thread launch determine the number of
|
threads we should spawn for this team. */
|
threads we should spawn for this team. */
|
/* ??? I have no idea what best practice for this is. Surely some
|
/* ??? I have no idea what best practice for this is. Surely some
|
function of the number of processors that are *still* online and
|
function of the number of processors that are *still* online and
|
the load average. Here I use the number of processors online
|
the load average. Here I use the number of processors online
|
minus the 15 minute load average. */
|
minus the 15 minute load average. */
|
|
|
unsigned
|
unsigned
|
gomp_dynamic_max_threads (void)
|
gomp_dynamic_max_threads (void)
|
{
|
{
|
unsigned n_onln, loadavg, nthreads_var = gomp_icv (false)->nthreads_var;
|
unsigned n_onln, loadavg, nthreads_var = gomp_icv (false)->nthreads_var;
|
|
|
n_onln = get_num_procs ();
|
n_onln = get_num_procs ();
|
if (n_onln > nthreads_var)
|
if (n_onln > nthreads_var)
|
n_onln = nthreads_var;
|
n_onln = nthreads_var;
|
|
|
loadavg = 0;
|
loadavg = 0;
|
#ifdef HAVE_GETLOADAVG
|
#ifdef HAVE_GETLOADAVG
|
{
|
{
|
double dloadavg[3];
|
double dloadavg[3];
|
if (getloadavg (dloadavg, 3) == 3)
|
if (getloadavg (dloadavg, 3) == 3)
|
{
|
{
|
/* Add 0.1 to get a kind of biased rounding. */
|
/* Add 0.1 to get a kind of biased rounding. */
|
loadavg = dloadavg[2] + 0.1;
|
loadavg = dloadavg[2] + 0.1;
|
}
|
}
|
}
|
}
|
#endif
|
#endif
|
|
|
if (loadavg >= n_onln)
|
if (loadavg >= n_onln)
|
return 1;
|
return 1;
|
else
|
else
|
return n_onln - loadavg;
|
return n_onln - loadavg;
|
}
|
}
|
|
|
int
|
int
|
omp_get_num_procs (void)
|
omp_get_num_procs (void)
|
{
|
{
|
return get_num_procs ();
|
return get_num_procs ();
|
}
|
}
|
|
|
ialias (omp_get_num_procs)
|
ialias (omp_get_num_procs)
|
|
|
