Skip to content
This repository was archived by the owner on Mar 28, 2023. It is now read-only.

Commit 9a23a34

Browse files
t4c1t4c1
committed
split native floating point tests
1 parent 55795d3 commit 9a23a34

36 files changed

+2596
-36
lines changed

‎SYCL/AtomicRef/add.cpp

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -147,8 +147,6 @@
147147
// RUN: %CPU_RUN_PLACEHOLDER %t.out
148148
// RUN: %ACC_RUN_PLACEHOLDER %t.out
149149

150-
#define SYCL_USE_NATIVE_FP_ATOMICS
151-
152150
#include "add.h"
153151

154152
int main() { add_test_all<access::address_space::global_space>(); }

‎SYCL/AtomicRef/add.h

Lines changed: 5 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -348,6 +348,7 @@ template <access::address_space space> void add_test_all() {
348348
}
349349

350350
add_test_orders_scopes<space, double>(q, N);
351+
#ifndef FP_TESTS_ONLY
351352
if constexpr (sizeof(long) == 8) {
352353
add_test_orders_scopes<space, long>(q, N);
353354
add_test_orders_scopes<space, unsigned long>(q, N);
@@ -359,9 +360,11 @@ template <access::address_space space> void add_test_all() {
359360
if constexpr (sizeof(char *) == 8) {
360361
add_test_orders_scopes<space, char *, ptrdiff_t>(q, N);
361362
}
363+
#endif
362364
#else
363-
add_test_orders_scopes<space, int>(q, N);
364365
add_test_orders_scopes<space, float>(q, N);
366+
#ifndef FP_TESTS_ONLY
367+
add_test_orders_scopes<space, int>(q, N);
365368
add_test_orders_scopes<space, unsigned int>(q, N);
366369
if constexpr (sizeof(long) == 4) {
367370
add_test_orders_scopes<space, long>(q, N);
@@ -370,6 +373,7 @@ template <access::address_space space> void add_test_all() {
370373
if constexpr (sizeof(char *) == 4) {
371374
add_test_orders_scopes<space, char *, ptrdiff_t>(q, N);
372375
}
376+
#endif
373377
#endif
374378

375379
std::cout << "Test passed." << std::endl;

‎SYCL/AtomicRef/add_generic.cpp

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -150,8 +150,6 @@
150150
// CUDA and HIP backends have had no support for the generic address space yet
151151
// XFAIL: cuda || hip
152152

153-
#define SYCL_USE_NATIVE_FP_ATOMICS
154-
155153
#include "add.h"
156154

157155
int main() { add_test_all<access::address_space::generic_space>(); }

‎SYCL/AtomicRef/add_generic_local.cpp

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -152,7 +152,6 @@
152152
// XFAIL: cuda || hip || host
153153

154154
#define TEST_GENERIC_IN_LOCAL 1
155-
#define SYCL_USE_NATIVE_FP_ATOMICS
156155

157156
#include "add.h"
158157

‎SYCL/AtomicRef/add_generic_local_native_fp.cpp

Lines changed: 163 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,163 @@
1+
// Each combination of 64/32 bit atomic, relaxed/acquire/release/acq_rel
2+
// semantic order and sub_group/work_group/device/system scope is tested
3+
// separately. This is controlled by macros, defined by RUN commands. Defaults
4+
// (no macro for a group) are: 32 bit, relaxed and device.
5+
6+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60
7+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
8+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
9+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
10+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
11+
12+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DATOMIC64
13+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
14+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
15+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
16+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
17+
18+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DWORK_GROUP
19+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
20+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
21+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
22+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
23+
24+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DWORK_GROUP -DATOMIC64
25+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
26+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
27+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
28+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
29+
30+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DSYSTEM
31+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
32+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
33+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
34+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
35+
36+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DSYSTEM -DATOMIC64
37+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
38+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
39+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
40+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
41+
42+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE
43+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
44+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
45+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
46+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
47+
48+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DATOMIC64
49+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
50+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
51+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
52+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
53+
54+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DWORK_GROUP
55+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
56+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
57+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
58+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
59+
60+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DWORK_GROUP -DATOMIC64
61+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
62+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
63+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
64+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
65+
66+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DSYSTEM
67+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
68+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
69+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
70+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
71+
72+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DSYSTEM -DATOMIC64
73+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
74+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
75+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
76+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
77+
78+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE
79+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
80+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
81+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
82+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
83+
84+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DATOMIC64
85+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
86+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
87+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
88+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
89+
90+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DWORK_GROUP
91+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
92+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
93+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
94+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
95+
96+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DWORK_GROUP -DATOMIC64
97+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
98+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
99+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
100+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
101+
102+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DSYSTEM
103+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
104+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
105+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
106+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
107+
108+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DSYSTEM -DATOMIC64
109+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
110+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
111+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
112+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
113+
114+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL
115+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
116+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
117+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
118+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
119+
120+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DATOMIC64
121+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
122+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
123+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
124+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
125+
126+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DWORK_GROUP
127+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
128+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
129+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
130+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
131+
132+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DWORK_GROUP -DATOMIC64
133+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
134+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
135+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
136+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
137+
138+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DSYSTEM
139+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
140+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
141+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
142+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
143+
144+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DSYSTEM -DATOMIC64
145+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
146+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
147+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
148+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
149+
150+
// CUDA and HIP backends have had no support for the generic address space yet.
151+
// Host does not support barrier.
152+
// XFAIL: cuda || hip || host
153+
154+
// hip, level zero and opencl do not support native floating point atomics
155+
// XFAIL: hip, level_zero, opencl
156+
157+
#define SYCL_USE_NATIVE_FP_ATOMICS
158+
#define FP_TESTS_ONLY
159+
#define TEST_GENERIC_IN_LOCAL 1
160+
161+
#include "add.h"
162+
163+
int main() { add_test_all<access::address_space::generic_space>(); }

‎SYCL/AtomicRef/add_generic_native_fp.cpp

Lines changed: 161 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,161 @@
1+
// Each combination of 64/32 bit atomic, relaxed/acquire/release/acq_rel
2+
// semantic order and sub_group/work_group/device/system scope is tested
3+
// separately. This is controlled by macros, defined by RUN commands. Defaults
4+
// (no macro for a group) are: 32 bit, relaxed and device.
5+
6+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60
7+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
8+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
9+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
10+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
11+
12+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DATOMIC64
13+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
14+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
15+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
16+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
17+
18+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DWORK_GROUP
19+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
20+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
21+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
22+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
23+
24+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DWORK_GROUP -DATOMIC64
25+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
26+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
27+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
28+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
29+
30+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DSYSTEM
31+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
32+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
33+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
34+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
35+
36+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 -DSYSTEM -DATOMIC64
37+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
38+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
39+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
40+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
41+
42+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE
43+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
44+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
45+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
46+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
47+
48+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DATOMIC64
49+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
50+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
51+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
52+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
53+
54+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DWORK_GROUP
55+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
56+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
57+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
58+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
59+
60+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DWORK_GROUP -DATOMIC64
61+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
62+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
63+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
64+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
65+
66+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DSYSTEM
67+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
68+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
69+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
70+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
71+
72+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQUIRE -DSYSTEM -DATOMIC64
73+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
74+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
75+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
76+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
77+
78+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE
79+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
80+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
81+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
82+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
83+
84+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DATOMIC64
85+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
86+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
87+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
88+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
89+
90+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DWORK_GROUP
91+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
92+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
93+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
94+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
95+
96+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DWORK_GROUP -DATOMIC64
97+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
98+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
99+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
100+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
101+
102+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DSYSTEM
103+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
104+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
105+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
106+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
107+
108+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DRELEASE -DSYSTEM -DATOMIC64
109+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
110+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
111+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
112+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
113+
114+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL
115+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
116+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
117+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
118+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
119+
120+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DATOMIC64
121+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
122+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
123+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
124+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
125+
126+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DWORK_GROUP
127+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
128+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
129+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
130+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
131+
132+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DWORK_GROUP -DATOMIC64
133+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
134+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
135+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
136+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
137+
138+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DSYSTEM
139+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
140+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
141+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
142+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
143+
144+
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 -DACQ_REL -DSYSTEM -DATOMIC64
145+
// RUN: %HOST_RUN_PLACEHOLDER %t.out
146+
// RUN: %GPU_RUN_PLACEHOLDER %t.out
147+
// RUN: %CPU_RUN_PLACEHOLDER %t.out
148+
// RUN: %ACC_RUN_PLACEHOLDER %t.out
149+
150+
// CUDA and HIP backends have had no support for the generic address space yet
151+
// XFAIL: cuda || hip
152+
153+
// hip, level zero and opencl do not support native floating point atomics
154+
// XFAIL: hip, level_zero, opencl
155+
156+
#define SYCL_USE_NATIVE_FP_ATOMICS
157+
#define FP_TESTS_ONLY
158+
159+
#include "add.h"
160+
161+
int main() { add_test_all<access::address_space::generic_space>(); }

‎SYCL/AtomicRef/add_local.cpp

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -151,8 +151,6 @@
151151
// point atomics.
152152
// XFAIL: host, hip, opencl
153153

154-
//#define SYCL_USE_NATIVE_FP_ATOMICS
155-
156154
#include "add.h"
157155

158156
int main() { add_test_all<access::address_space::local_space>(); }

0 commit comments

Comments
 (0)