1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
|
subroutine initialize_array(n, a)
use omp_lib
implicit none
integer, intent(in) :: n
real, dimension(n), intent(out) :: a
integer :: i
!$omp parallel
!$omp do
do i = 1, n
a(i) = 12.91
end do
!$omp end do
!$omp end parallel
print *, a(1), a(n)
end subroutine
subroutine parallel_sum(n, a, ctr)
use omp_lib
implicit none
integer, intent(in) :: n, ctr
real, dimension(n), intent(in) :: a
integer :: i
real :: sum
sum = 0.0
!$omp parallel do reduction(+:sum)
do i = 1, n
sum = sum + a(i)
end do
!$omp end parallel do
! here we have precision error both with gfortran and lfortran
! setting different number of threads leads to different results
! but in any case output of gfortran and lfortran is the same
print *, sum
if ( ctr == 1 ) then
if (abs(sum - 12933994.0) > 1e-7) error stop
else
! if (abs(sum - 25867988.0) > 1e-7) error stop ! lfortran always gives 2.58679600e+07
end if
end subroutine
subroutine reduce_array(n, a, b)
implicit none
integer, intent(in) :: n
real, dimension(n), intent(in) :: a
real, dimension(n) :: b
integer :: i
!$omp parallel do
do i = 2, n
b(i) = a(i) + a(i-1)
end do
!$omp end parallel do
b(1) = a(1) * 2
end subroutine
program openmp_15
use omp_lib
integer, parameter :: n = 1000000
real, dimension(n) :: a, b
call omp_set_num_threads(4)
call initialize_array(n, a)
print *, a(1), a(n)
call parallel_sum(n, a, 1)
call reduce_array(n, a, b)
call parallel_sum(n, b, 0)
end program
|
