| 1 |
! { dg-do run } |
| 2 |
! { dg-additional-options "-msse2" { target sse2_runtime } } |
| 3 |
! { dg-additional-options "-mavx" { target avx_runtime } } |
| 4 |
|
| 5 |
integer :: a(1024), b(1024), k, m, i, s, t |
| 6 |
k = 4 |
| 7 |
m = 2 |
| 8 |
t = 1 |
| 9 |
do i = 1, 1024 |
| 10 |
a(i) = i - 513 |
| 11 |
b(i) = modulo (i - 52, 39) |
| 12 |
if (i.lt.52.and.b(i).ne.0) b(i) = b(i) - 39 |
| 13 |
end do |
| 14 |
s = foo (b) |
| 15 |
do i = 1, 1024 |
| 16 |
if (a(i).ne.((i - 513) * b(i))) stop 1 |
| 17 |
if (i.lt.52.and.modulo (i - 52, 39).ne.0) then |
| 18 |
if (b(i).ne.(modulo (i - 52, 39) - 39)) stop 2 |
| 19 |
else |
| 20 |
if (b(i).ne.(modulo (i - 52, 39))) stop 3 |
| 21 |
end if |
| 22 |
a(i) = i - 513 |
| 23 |
end do |
| 24 |
if (k.ne.(4 + 3 * 1024).or.s.ne.1596127) stop 4 |
| 25 |
k = 4 |
| 26 |
m = 2 |
| 27 |
t = 1 |
| 28 |
s = bar (b) |
| 29 |
do i = 1, 1024 |
| 30 |
if (a(i).ne.((i - 513) * b(i))) stop 5 |
| 31 |
if (i.lt.52.and.modulo (i - 52, 39).ne.0) then |
| 32 |
if (b(i).ne.(modulo (i - 52, 39) - 39)) stop 6 |
| 33 |
else |
| 34 |
if (b(i).ne.(modulo (i - 52, 39))) stop 7 |
| 35 |
end if |
| 36 |
a(i) = i - 513 |
| 37 |
end do |
| 38 |
if (k.ne.(4 + 3 * 1024).or.s.ne.1596127) stop 8 |
| 39 |
k = 4 |
| 40 |
m = 2 |
| 41 |
t = 1 |
| 42 |
s = baz (b) |
| 43 |
do i = 1, 1024 |
| 44 |
if (a(i).ne.((i - 513) * b(i))) stop 9 |
| 45 |
if (i.lt.52.and.modulo (i - 52, 39).ne.0) then |
| 46 |
if (b(i).ne.(modulo (i - 52, 39) - 39)) stop 10 |
| 47 |
else |
| 48 |
if (b(i).ne.(modulo (i - 52, 39))) stop 11 |
| 49 |
end if |
| 50 |
end do |
| 51 |
if (k.ne.(4 + 3 * 1024).or.s.ne.1596127) stop 12 |
| 52 |
contains |
| 53 |
function foo (p) |
| 54 |
integer :: p(1024), u, v, i, s, foo |
| 55 |
s = 0 |
| 56 |
!$omp parallel do simd linear(k : m + 1) reduction(+: s) & |
| 57 |
!$omp & lastprivate(u, v) schedule (static, 32) |
| 58 |
do i = 1, 1024 |
| 59 |
a(i) = a(i) * p(i) |
| 60 |
u = p(i) + k |
| 61 |
k = k + m + 1 |
| 62 |
v = p(i) + k |
| 63 |
s = s + p(i) + k |
| 64 |
end do |
| 65 |
!$omp end parallel do simd |
| 66 |
if (i.ne.1025) stop 13 |
| 67 |
if (u.ne.(36 + 4 + 3 * 1023).or.v.ne.(36 + 4 + 3 * 1024)) stop 14 |
| 68 |
foo = s |
| 69 |
end function foo |
| 70 |
function bar (p) |
| 71 |
integer :: p(1024), u, v, i, s, bar |
| 72 |
s = 0 |
| 73 |
!$omp parallel do simd linear(k : m + 1) reduction(+: s) & |
| 74 |
!$omp & lastprivate(u, v) schedule (dynamic, 32) |
| 75 |
do i = 1, 1024, t |
| 76 |
a(i) = a(i) * p(i) |
| 77 |
u = p(i) + k |
| 78 |
k = k + m + 1 |
| 79 |
v = p(i) + k |
| 80 |
s = s + p(i) + k |
| 81 |
end do |
| 82 |
!$omp endparalleldosimd |
| 83 |
if (i.ne.1025) stop 15 |
| 84 |
if (u.ne.(36 + 4 + 3 * 1023).or.v.ne.(36 + 4 + 3 * 1024)) stop 16 |
| 85 |
bar = s |
| 86 |
end function bar |
| 87 |
function baz (p) |
| 88 |
integer :: p(1024), u, v, i, s, baz |
| 89 |
s = 0 |
| 90 |
!$omp parallel do simd linear(k : m + 1) reduction(+: s) & |
| 91 |
!$omp & lastprivate(u, v) linear(i : t) schedule (static, 8) |
| 92 |
do i = 1, 1024, t |
| 93 |
a(i) = a(i) * p(i) |
| 94 |
u = p(i) + k |
| 95 |
k = k + m + 1 |
| 96 |
v = p(i) + k |
| 97 |
s = s + p(i) + k |
| 98 |
end do |
| 99 |
if (i.ne.1025) stop 17 |
| 100 |
if (u.ne.(36 + 4 + 3 * 1023).or.v.ne.(36 + 4 + 3 * 1024)) stop 18 |
| 101 |
baz = s |
| 102 |
end function baz |
| 103 |
end |