blob: 72d7af6dfd414458b148f37ec422eaba9a61cfa6 (
plain)
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
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
|
/* { dg-do run } */
/* { dg-set-target-env-var OMP_NUM_TEAMS_ALL "3" } */
/* { dg-set-target-env-var OMP_NUM_TEAMS_DEV "4" } */
/* { dg-set-target-env-var OMP_NUM_TEAMS "5" } */
/* { dg-set-target-env-var OMP_NUM_TEAMS_DEV_0 "6" } */
/* { dg-set-target-env-var OMP_NUM_TEAMS_DEV_1 "7" } */
/* { dg-set-target-env-var OMP_NUM_TEAMS_DEV_2 "8" } */
/* { dg-set-target-env-var OMP_TEAMS_THREAD_LIMIT_ALL "2" } */
/* { dg-set-target-env-var OMP_TEAMS_THREAD_LIMIT_DEV "3" } */
/* { dg-set-target-env-var OMP_TEAMS_THREAD_LIMIT "4" } */
/* { dg-set-target-env-var OMP_TEAMS_THREAD_LIMIT_DEV_0 "5" } */
/* { dg-set-target-env-var OMP_TEAMS_THREAD_LIMIT_DEV_1 "6" } */
/* { dg-set-target-env-var OMP_TEAMS_THREAD_LIMIT_DEV_2 "7" } */
#include <omp.h>
#include <stdlib.h>
#include <unistd.h>
int
main ()
{
if (omp_get_max_teams () != 5
|| omp_get_teams_thread_limit () != 4)
abort ();
#pragma omp teams
{
if (omp_get_num_teams () > 5
|| omp_get_team_num () >= 5)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 4
|| omp_get_thread_num () >= 4)
abort ();
}
omp_set_num_teams (4);
omp_set_teams_thread_limit (3);
if (omp_get_max_teams () != 4
|| omp_get_teams_thread_limit () != 3)
abort ();
#pragma omp teams
{
if (omp_get_num_teams () > 4
|| omp_get_team_num () >= 4)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 3
|| omp_get_thread_num () >= 3)
abort ();
}
#pragma omp teams num_teams(3) thread_limit(2)
{
if (omp_get_num_teams () != 3
|| omp_get_team_num () >= 3)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 2
|| omp_get_thread_num () >= 2)
abort ();
}
#pragma omp teams num_teams(5) thread_limit(4)
{
if (omp_get_num_teams () != 5
|| omp_get_team_num () >= 5)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 4
|| omp_get_thread_num () >= 4)
abort ();
}
int num_devices = omp_get_num_devices () > 3 ? 3 : omp_get_num_devices ();
for (int i = 0; i < num_devices; i++)
{
#pragma omp target device (i)
if (omp_get_max_teams () != 6 + i
|| omp_get_teams_thread_limit () != 5 + i)
abort ();
#pragma omp target device (i)
#pragma omp teams
#pragma omp parallel
if (omp_get_thread_limit () > 5 + i
|| omp_get_thread_num () >= 5 + i)
abort ();
#pragma omp target device (i)
{
omp_set_num_teams (5 + i);
omp_set_teams_thread_limit (4 + i);
if (omp_get_max_teams () != 5 + i
|| omp_get_teams_thread_limit () != 4 + i)
abort ();
}
/* omp_set_num_teams and omp_set_teams_thread_limit above set the value
of nteams-var and teams-thread-limit-var ICVs on device 'i', which has
scope 'device' and should be avaible in subsequent target regions. */
#pragma omp target device (i)
if (omp_get_max_teams () != 5 + i
|| omp_get_teams_thread_limit () != 4 + i)
abort ();
#pragma omp target device (i)
#pragma omp teams
{
if (omp_get_num_teams () > 5 + i
|| omp_get_team_num () >= 5 + i)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 4 + i
|| omp_get_thread_num () >= 4 + i)
abort ();
}
#pragma omp target device (i)
#pragma omp teams num_teams(6 + i) thread_limit(5 + i)
{
if (omp_get_num_teams () > 6 + i
|| omp_get_team_num () >= 6 + i)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 5 + i
|| omp_get_thread_num () >= 5 + i
|| omp_get_num_teams () > 6 + i
|| omp_get_team_num () >= 6 + i)
abort ();
}
#pragma omp target device (i)
#pragma omp teams num_teams(4 + i) thread_limit(3 + i)
{
if (omp_get_num_teams () > 4 + i
|| omp_get_team_num () >= 4 + i)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 3 + i
|| omp_get_thread_num () >= 3 + i
|| omp_get_num_teams () > 4 + i
|| omp_get_team_num () >= 4 + i)
abort ();
}
#pragma omp target device (i)
#pragma omp teams thread_limit(3 + i) num_teams(4 + i)
{
if (omp_get_num_teams () > 4 + i
|| omp_get_team_num () >= 4 + i)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > 3 + i
|| omp_get_thread_num () >= 3 + i
|| omp_get_num_teams () > 4 + i
|| omp_get_team_num () >= 4 + i)
abort ();
}
/* The NUM_TEAMS and THREAD_LIMIT clauses should not change the values
of the corresponding ICVs. */
#pragma omp target device (i)
if (omp_get_max_teams () != 5 + i
|| omp_get_teams_thread_limit () != 4 + i)
abort ();
/* This tests a large number of teams and threads. If it is larger than
2^15+1 then the according argument in the kernels arguments list
is encoded with two items instead of one. */
intptr_t large_num_teams = 66000;
intptr_t large_threads_limit = 67000;
#pragma omp target device (i)
{
omp_set_num_teams (large_num_teams + i);
omp_set_teams_thread_limit (large_threads_limit + i);
if (omp_get_max_teams () != large_num_teams + i
|| omp_get_teams_thread_limit () != large_threads_limit + i)
abort ();
}
#pragma omp target device (i)
if (omp_get_max_teams () != large_num_teams + i
|| omp_get_teams_thread_limit () != large_threads_limit + i)
abort ();
#pragma omp target device (i)
#pragma omp teams
{
if (omp_get_num_teams () > large_num_teams + i
|| omp_get_team_num () >= large_num_teams + i)
abort ();
#pragma omp parallel
if (omp_get_thread_limit () > large_threads_limit + i
|| omp_get_thread_num () >= large_threads_limit + i)
abort ();
}
}
return 0;
}
|
