OpenMP Application Programming Interface Examples

1
2
3
4
5
6
1.16 The omp_get_num_threads Routine
In the following example, the omp_get_num_threads call returns 1 in the sequential part of
the code, so np will always be equal to 1. To determine the number of threads that will be deployed
for the parallel region, the call should be inside the parallel region.
Example get_nthrs.1.c
C / C++
S-1 #include
S-2 void work(int i);
S-3
S-4 void incorrect()
S-5 {
S-6 int np, i;
S-7
S-8 np = omp_get_num_threads(); /* misplaced */
S-9
S-10 #pragma omp parallel for schedule(static)
S-11 for (i=0; i < np; i++)
S-12 work(i);
S-13 }
C / C++
Fortran
Example get_nthrs.1.f
S-1 SUBROUTINE WORK(I)
S-2 INTEGER I
S-3 I = I + 1
S-4 END SUBROUTINE WORK
S-5
S-6 SUBROUTINE INCORRECT()
S-7 INCLUDE "omp_lib.h" ! or USE OMP_LIB
S-8 INTEGER I, NP
S-9
S-10 NP = OMP_GET_NUM_THREADS() !misplaced: will return 1
S-11 !$OMP PARALLEL DO SCHEDULE(STATIC)
S-12 DO I = 0, NP-1
S-13 CALL WORK(I)
S-14 ENDDO
S-15 !$OMP END PARALLEL DO
S-16 END SUBROUTINE INCORRECT
38 OpenMP Examples Version 4.5.0 - November 2016