Skip to content

Commit a02513c

Browse files
whitneywhtsangwhitneywhtsang
committed
[Gluon] Mark target specific tests as xfail
Signed-off-by: Whitney Tsang <[email protected]>
1 parent 4e5de6b commit a02513c

File tree

1 file changed

+19
-17
lines changed

1 file changed

+19
-17
lines changed

‎python/test/gluon/test_consan.py‎

Lines changed: 19 additions & 17 deletions
Original file line numberDiff line numberDiff line change
@@ -639,7 +639,7 @@ def kernel(input_desc, FAILURE: ttgl.constexpr):
639639
kernel[(1, )](input_desc, FAILURE=FAILURE, num_warps=4)
640640

641641

642-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
642+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
643643
@pytest.mark.parametrize("FAILURE", [True, False])
644644
def test_ws_store_wait_load(FAILURE, device, run_wrapper):
645645
if run_wrapper:
@@ -692,7 +692,7 @@ def ws_kernel(output, FAILURE: ttgl.constexpr):
692692
ws_kernel[(1, )](output, FAILURE=FAILURE, num_warps=4)
693693

694694

695-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
695+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
696696
@pytest.mark.parametrize("FAILURE", [True, False])
697697
def test_ws_load_wait_store(FAILURE, device, run_wrapper):
698698
if run_wrapper:
@@ -745,7 +745,7 @@ def ws_kernel(output, FAILURE: ttgl.constexpr):
745745
ws_kernel[(1, )](output, FAILURE=FAILURE, num_warps=4)
746746

747747

748-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
748+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
749749
@pytest.mark.parametrize("MISSING_BAR", ["none", "1", "2"])
750750
def test_ws_two_loads_two_bars(MISSING_BAR, device, run_wrapper):
751751
if run_wrapper:
@@ -807,7 +807,7 @@ def kernel(output, MISSING_BAR: ttgl.constexpr):
807807
kernel[(1, )](output, MISSING_BAR=MISSING_BAR, num_warps=4)
808808

809809

810-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
810+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
811811
@pytest.mark.parametrize("FAILURE", [True, False])
812812
def test_ws_two_loads_one_bar(FAILURE, device, run_wrapper):
813813
if run_wrapper:
@@ -866,7 +866,7 @@ def kernel(output, FAILURE: ttgl.constexpr):
866866
kernel[(1, )](output, FAILURE=FAILURE, num_warps=4)
867867

868868

869-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
869+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
870870
@pytest.mark.parametrize("MISSING_BAR", ["none", "0", "1", "2", "3"])
871871
def test_ws_two_loads_two_bars_loop(MISSING_BAR, device, run_wrapper):
872872
if run_wrapper:
@@ -949,7 +949,7 @@ def kernel(output, MISSING_BAR: ttgl.constexpr):
949949
kernel[(1, )](output, MISSING_BAR=MISSING_BAR, num_warps=4)
950950

951951

952-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
952+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
953953
@pytest.mark.parametrize("FAILURE", [True, False])
954954
def test_ws_load_ordering(FAILURE, device, run_wrapper):
955955
if run_wrapper:
@@ -1013,7 +1013,7 @@ def kernel(output, FAILURE: ttgl.constexpr):
10131013
kernel[(1, )](output, FAILURE=FAILURE, num_warps=4)
10141014

10151015

1016-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1016+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
10171017
@pytest.mark.parametrize("MISSING_BAR", ["none", "T2", "T3"])
10181018
def test_ws_two_producers_two_consumers(MISSING_BAR, device, run_wrapper):
10191019
if run_wrapper:
@@ -1100,7 +1100,7 @@ def kernel(output, MISSING_BAR: ttgl.constexpr):
11001100
kernel[(1, )](output, MISSING_BAR=MISSING_BAR, num_warps=4)
11011101

11021102

1103-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1103+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
11041104
@pytest.mark.parametrize("MISSING_BAR", ["none", "1", "2"])
11051105
def test_ws_different_warp_sizes(MISSING_BAR, device, run_wrapper):
11061106
if run_wrapper:
@@ -1168,7 +1168,8 @@ def kernel(output, MISSING_BAR: ttgl.constexpr):
11681168
kernel[(1, )](output, MISSING_BAR=MISSING_BAR, num_warps=4)
11691169

11701170

1171-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper or newer")
1171+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper or newer",
1172+
run=False)
11721173
@pytest.mark.parametrize("FAILURE", [True, False])
11731174
def test_ws_async_copy_commits(FAILURE, device, run_wrapper):
11741175
if run_wrapper:
@@ -1229,7 +1230,8 @@ def kernel(input, FAILURE: ttgl.constexpr):
12291230
kernel[(1, )](input, FAILURE=FAILURE, num_warps=4)
12301231

12311232

1232-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper or newer")
1233+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper or newer",
1234+
run=False)
12331235
@pytest.mark.parametrize("FAILURE", [True, False])
12341236
def test_ws_async_copy_wait_visibility(FAILURE, device, run_wrapper):
12351237
if run_wrapper:
@@ -1283,7 +1285,7 @@ def kernel(input, FAILURE: ttgl.constexpr):
12831285
kernel[(1, )](input, FAILURE=FAILURE, num_warps=4)
12841286

12851287

1286-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] != 9, reason="Requires hopper")
1288+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] != 9, reason="Requires hopper", run=False)
12871289
@pytest.mark.parametrize("FAILURE", [True, False])
12881290
def test_ws_wgmma_wait_visibility(FAILURE, device, run_wrapper):
12891291
if run_wrapper:
@@ -1337,7 +1339,7 @@ def kernel(FAILURE: ttgl.constexpr):
13371339
kernel[(1, )](FAILURE=FAILURE, num_warps=4)
13381340

13391341

1340-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1342+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
13411343
def test_deadlock_two_partitions(device, run_wrapper):
13421344
if run_wrapper:
13431345
result = run_in_process(test_deadlock_two_partitions, (device, False))
@@ -1371,7 +1373,7 @@ def kernel():
13711373
kernel[(1, )](num_warps=4)
13721374

13731375

1374-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1376+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
13751377
def test_deadlock_overarrival(device, run_wrapper):
13761378
if run_wrapper:
13771379
result = run_in_process(test_deadlock_overarrival, (device, False))
@@ -1400,7 +1402,7 @@ def kernel():
14001402
kernel[(1, )](num_warps=4)
14011403

14021404

1403-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1405+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
14041406
def test_deadlock_underarrival(device, run_wrapper):
14051407
if run_wrapper:
14061408
result = run_in_process(test_deadlock_underarrival, (device, False))
@@ -1436,7 +1438,7 @@ def kernel():
14361438
kernel[(1, )](num_warps=4)
14371439

14381440

1439-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1441+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
14401442
def test_deadlock_different_phases(device, run_wrapper):
14411443
if run_wrapper:
14421444
result = run_in_process(test_deadlock_different_phases, (device, False))
@@ -1471,7 +1473,7 @@ def kernel():
14711473
kernel[(1, )](num_warps=4)
14721474

14731475

1474-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1476+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
14751477
def test_deadlock_exempt_when_tma_signals(device, run_wrapper):
14761478
if run_wrapper:
14771479
result = run_in_process(test_deadlock_exempt_when_tma_signals, (device, False))
@@ -1514,7 +1516,7 @@ def kernel(input_desc):
15141516
kernel[(1, )](input_desc, num_warps=4)
15151517

15161518

1517-
@pytest.mark.skipif(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper")
1519+
@pytest.mark.xfail(not is_cuda() or torch.cuda.get_device_capability()[0] < 9, reason="Requires hopper", run=False)
15181520
def test_barrier_underflow(device, run_wrapper):
15191521
if run_wrapper:
15201522
result = run_in_process(test_barrier_underflow, (device, False))

0 commit comments

Comments
 (0)