Memory leak or incorrect use of '#pragma omp parallel'?

[Mikolaj Golub]

Hello,

If I run the following program

---

#include 

int n = 4, m = 2;

int main () {
for (;;) {
int i;

#pragma omp parallel num_threads(m)
{
int just_to_make_some_code_generated;
just_to_make_some_code_generated = 0;
}
#pragma omp parallel for num_threads(n)
for(i = 0; i < 1; i++) {}

}

return 0;
}

---

on FreeBSD6/7 (i386 and amd64) built with gcc42 or gcc43 -fopenmp (I have not
tried other gcc versions), the memory usage is constantly growing during the
run. The memory growth is observed in case n != m. If n == m, everything works
ok.

I have tried this code on CentOS5.2 i386 built with gcc41 and gcc43 and
constant memory growth is not observed here. Althogh memory usage looks a bit
strange. On the start the program uses much more memory (more then 60Mb) in
comparison with FreeBSD (several Mb). If n == m, the memory usage statistis is
not changing during the run, while when n != m the memory usage statistis
varies from 64Mb to 102Mb.

Should I report this as a bug? I am not familiar with openmp, we faced with
this problem running some real application. So I am not sure if this is gcc
bug or just incorrect use of '#pragma omp parallel'.

-- 
Mikolaj Golub

Re: Memory leak or incorrect use of '#pragma omp parallel'?

[Mikolaj Golub]

On Mon, 11 May 2009 19:00:54 +0300 Mikolaj Golub wrote:

> Hello,
>
> If I run the following program
>
> ---
>
> #include 
>
> int n = 4, m = 2;
>
> int main () {
> for (;;) {
> int i;
>
> #pragma omp parallel num_threads(m)
> {
> int just_to_make_some_code_generated;
> just_to_make_some_code_generated = 0;
> }
> #pragma omp parallel for num_threads(n)
> for(i = 0; i < 1; i++) {}
>
> }
>
> return 0;
> }
>
> ---
>
> on FreeBSD6/7 (i386 and amd64) built with gcc42 or gcc43 -fopenmp (I have not
> tried other gcc versions), the memory usage is constantly growing during the
> run. The memory growth is observed in case n != m. If n == m, everything works
> ok.

The problem is observed also for constructions like this one:

#pragma omp parallel for num_threads(m)
...
#pragma omp parallel for num_threads(n)
...

Adding some sleep() code I see in top that during the program run m or n
threads are run in turn. So it looks like memory leaks when thread is
removed. It might be FreeBSD problem, not gcc.

-- 
Mikolaj Golub

Re: Memory leak or incorrect use of '#pragma omp parallel'?

[Mikolaj Golub]

On Tue, 12 May 2009 09:33:34 +0300 Mikolaj Golub wrote:

 MG> On Mon, 11 May 2009 19:00:54 +0300 Mikolaj Golub wrote:

 >> Hello,
 >>
 >> If I run the following program
 >>
 >> ---
 >>
 >> #include 
 >>
 >> int n = 4, m = 2;
 >>
 >> int main () {
 >> for (;;) {
 >> int i;
 >>
 >> #pragma omp parallel num_threads(m)
 >> {
 >> int just_to_make_some_code_generated;
 >> just_to_make_some_code_generated = 0;
 >> }
 >> #pragma omp parallel for num_threads(n)
 >> for(i = 0; i < 1; i++) {}
 >>
 >> }
 >>
 >> return 0;
 >> }
 >>
 >> ---
 >>
 >> on FreeBSD6/7 (i386 and amd64) built with gcc42 or gcc43 -fopenmp (I have 
 >> not
 >> tried other gcc versions), the memory usage is constantly growing during the
 >> run. The memory growth is observed in case n != m. If n == m, everything 
 >> works
 >> ok.

 MG> The problem is observed also for constructions like this one:

 MG> #pragma omp parallel for num_threads(m)
 MG> ...
 MG> #pragma omp parallel for num_threads(n)
 MG> ...

 MG> Adding some sleep() code I see in top that during the program run m or n
 MG> threads are run in turn. So it looks like memory leaks when thread is
 MG> removed. It might be FreeBSD problem, not gcc.

The problem is in libgomp/team.c.  gomp_thread_start() does gomp_sem_init()
but gomp_sem_destroy() is never called. FreeBSD implementation of sem_init()
allocates some memory for semaphore. This patch solves the problem for me:

--- gcc-4.4-20090227/libgomp/team.c.orig2009-05-16 22:57:05.0 
+0300
+++ gcc-4.4-20090227/libgomp/team.c 2009-05-16 22:57:14.0 +0300
@@ -458,6 +458,7 @@ gomp_team_start (void (*fn) (void *), vo
 void
 gomp_team_end (void)
 {
+  int i;
   struct gomp_thread *thr = gomp_thread ();
   struct gomp_team *team = thr->ts.team;
 
@@ -493,6 +494,9 @@ gomp_team_end (void)
}
   while (ws != NULL);
 }
+
+  for(i = 1; i < team->nthreads; i++)
+gomp_sem_destroy (team->ordered_release[i]);
   gomp_sem_destroy (&team->master_release);
 #ifndef HAVE_SYNC_BUILTINS
   gomp_mutex_destroy (&team->work_share_list_free_lock);

I am going to register this in bugzilla.

The problem is not observed under Linux, because it looks like sem_init()
implementation for Linux does not do memory allocation in the process' virtual
space. The memory for the test program below grows under FreeBSD and does not
under Linux.

#include 

int
main(int argc, char *argv[]) {

sem_t sem;

for(;;) { sem_init(&sem, 0, 0);}

return 0;
}

-- 
Mikolaj Golub

Re: Memory leak or incorrect use of '#pragma omp parallel'?

[Mikolaj Golub]

On Sun, 17 May 2009 09:34:22 +0300 Mikolaj Golub wrote:

 MG> The problem is in libgomp/team.c.  gomp_thread_start() does gomp_sem_init()
 MG> but gomp_sem_destroy() is never called.

Registered in bugzilla

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40174

-- 
Mikolaj Golub