View Source

{section:border=true}
{column:width=70%}

h2. AUTO Loop Schedule in OpenMP 3.0

The AUTO loop schedule is a new loop schedule introduced in OpenMP 3.0. When SCHEDULE(AUTO) is specified on a do/for loop directive, the decision regarding scheduling is delegated to the compiler and/or OpenMP runtime support library. The programmer gives the compiler and/or libmtsk the freedom to choose any possible mapping of iterations to threads in the team.

The AUTO loop schedule is described in Section 2.5.1 (p. 38) of the OpenMP specification.

*Example (Fortran):*
{noformat}
Program Test
integer A(10), B(10)

A = -1
B = 20

!$OMP PARALLEL DO PRIVATE (A), SHARED(B), SCHEDULE(AUTO)
do i = 1, 10
A(i) = i
B(i) = B(i) + A(i)
end do

do i = 1, 10
PRINT *, B(i)
end do
end


% f90 -xopenmp -xO3 test.f
% a.out
21
22
23
24
25
26
27
28
29
30
{noformat}

*Example (C/C++):*
{noformat}
#include <stdio.h>
#include <omp.h>

#define SIZE 10

int main(void)
{
int i;
int buffer[SIZE];

#pragma omp parallel for schedule(auto)
for (i = 0; i < SIZE; i++)
{
buffer[i] = i;
}

for (i = 0; i < SIZE; i++)
{
printf ("%d\n", buffer[i]);
}

return 0;
}


% cc -xopenmp -xO3 test.c
% a.out
0
1
2
3
4
5
6
7
8
9
{noformat}
{column}
{column:width=30%}
{panel:title=OpenMP 3.0 Features in Express 7.08}
{children:page=Sun Studio OpenMP}
{panel}
{column}
{section}