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

Commit ddeac06

Browse files
t4c1t4c1
committed
speedup tests
1 parent 2ad0979 commit ddeac06

File tree

10 files changed

+33
-631
lines changed

10 files changed

+33
-631
lines changed

‎SYCL/AtomicRef/load.cpp

Lines changed: 1 addition & 72 deletions
Original file line numberDiff line numberDiff line change
@@ -1,75 +1,4 @@
1-
// Each combination of 64/32 bit atomic, relaxed/acquire
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_70 -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_70 -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
1+
// 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
732
// RUN: %HOST_RUN_PLACEHOLDER %t.out
743
// RUN: %GPU_RUN_PLACEHOLDER %t.out
754
// RUN: %CPU_RUN_PLACEHOLDER %t.out

‎SYCL/AtomicRef/load.h

Lines changed: 12 additions & 27 deletions
Original file line numberDiff line numberDiff line change
@@ -120,53 +120,37 @@ template <access::address_space space, typename T,
120120
void load_test_scopes(queue q, size_t N) {
121121
std::vector<memory_scope> scopes =
122122
q.get_device().get_info<info::device::atomic_memory_scope_capabilities>();
123-
#if defined(SYSTEM)
124-
if (std::find(scopes.begin(), scopes.end(), memory_scope::system) ==
123+
if (std::find(scopes.begin(), scopes.end(), memory_scope::system) !=
125124
scopes.end()) {
126-
std::cout << "Skipping cross device test\n";
127-
return;
125+
load_test<space, T, order, memory_scope::system>(q, N);
128126
}
129-
load_test<space, T, order, memory_scope::system>(q, N);
130-
#elif defined(WORK_GROUP)
131-
if (std::find(scopes.begin(), scopes.end(), memory_scope::system) ==
127+
if (std::find(scopes.begin(), scopes.end(), memory_scope::work_group) !=
132128
scopes.end()) {
133-
std::cout << "Skipping work group test\n";
134-
return;
129+
load_test<space, T, order, memory_scope::work_group>(q, N);
135130
}
136-
load_test<space, T, order, memory_scope::work_group>(q, N);
137-
#elif defined(SUB_GROUP)
138-
if (std::find(scopes.begin(), scopes.end(), memory_scope::system) ==
131+
if (std::find(scopes.begin(), scopes.end(), memory_scope::sub_group) !=
139132
scopes.end()) {
140-
std::cout << "Skipping sub group test\n";
141-
return;
133+
load_test<space, T, order, memory_scope::sub_group>(q, N);
142134
}
143-
load_test<space, T, order, memory_scope::sub_group>(q, N);
144-
#else
145135
load_test<space, T, order, memory_scope::device>(q, N);
146-
#endif
147136
}
148137

149138
template <access::address_space space, typename T>
150139
void load_test_orders_scopes(queue q, size_t N) {
151140
std::vector<memory_order> orders =
152141
q.get_device().get_info<info::device::atomic_memory_order_capabilities>();
153-
#if defined(ACQUIRE)
154-
if (std::find(orders.begin(), orders.end(), memory_order::acquire) ==
142+
if (std::find(orders.begin(), orders.end(), memory_order::acquire) !=
155143
orders.end()) {
156-
std::cout << "Skipping acquire tests\n";
157-
return;
144+
load_test_scopes<space, T, memory_order::acquire>(q, N);
158145
}
159-
load_test_scopes<space, T, memory_order::acquire>(q, N);
160-
#else
161146
load_test_scopes<space, T, memory_order::relaxed>(q, N);
162-
#endif
163147
}
164148

165149
template <access::address_space space> void load_test_all() {
166150
queue q;
167151

168152
constexpr int N = 32;
169-
#ifdef ATOMIC64
153+
#ifdef FULL_ATOMIC64_COVERAGE
170154
if (!q.get_device().has(aspect::atomic64)) {
171155
std::cout << "Skipping atomic64 tests\n";
172156
return;
@@ -184,9 +168,10 @@ template <access::address_space space> void load_test_all() {
184168
if constexpr (sizeof(char *) == 8) {
185169
load_test_orders_scopes<space, char *>(q, N);
186170
}
187-
#else
188-
load_test_orders_scopes<space, int>(q, N);
171+
#endif
189172
load_test_orders_scopes<space, float>(q, N);
173+
#ifdef FULL_ATOMIC32_COVERAGE
174+
load_test_orders_scopes<space, int>(q, N);
190175
load_test_orders_scopes<space, unsigned int>(q, N);
191176
if constexpr (sizeof(long) == 4) {
192177
load_test_orders_scopes<space, long>(q, N);

‎SYCL/AtomicRef/load_generic.cpp

Lines changed: 1 addition & 72 deletions
Original file line numberDiff line numberDiff line change
@@ -1,75 +1,4 @@
1-
// Each combination of 64/32 bit atomic, relaxed/acquire
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_70 -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_70 -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
1+
// 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
732
// RUN: %HOST_RUN_PLACEHOLDER %t.out
743
// RUN: %GPU_RUN_PLACEHOLDER %t.out
754
// RUN: %CPU_RUN_PLACEHOLDER %t.out

‎SYCL/AtomicRef/load_generic_local.cpp

Lines changed: 1 addition & 72 deletions
Original file line numberDiff line numberDiff line change
@@ -1,75 +1,4 @@
1-
// Each combination of 64/32 bit atomic, relaxed/acquire
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_70 -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_70 -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
1+
// 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
732
// RUN: %HOST_RUN_PLACEHOLDER %t.out
743
// RUN: %GPU_RUN_PLACEHOLDER %t.out
754
// RUN: %CPU_RUN_PLACEHOLDER %t.out

‎SYCL/AtomicRef/load_local.cpp

Lines changed: 1 addition & 72 deletions
Original file line numberDiff line numberDiff line change
@@ -1,75 +1,4 @@
1-
// Each combination of 64/32 bit atomic, relaxed/acquire
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_70 -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_70 -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
1+
// 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
732
// RUN: %HOST_RUN_PLACEHOLDER %t.out
743
// RUN: %GPU_RUN_PLACEHOLDER %t.out
754
// RUN: %CPU_RUN_PLACEHOLDER %t.out

0 commit comments

Comments
 (0)