Skip TestAutograd.test_profiler temporarily (#732)

We enabled PTI for Kenito Profiler XPU backend. The case fails. We retrieve the case once the bug is fixed. Tracking: #731 --------- Signed-off-by: Feng Yuan <[email protected]>
intel · Aug 10, 2024 · ab9feb9 · ab9feb9
1 parent eb14e58
commit ab9feb9
Showing 1 changed file with 12 additions and 0 deletions.
diff --git a/test/xpu/run_test_with_skip.py b/test/xpu/run_test_with_skip.py
@@ -1418,6 +1418,12 @@ def launch_test(test_case, skip_list=None, exe_list=None):
     "test_copy__xpu",
     "test_checkpointing_non_reentrant_autocast_cpu",
     "test_per_dispatch_key_input_saving_xpu",
+
+    # Runtime error after enabling PTI
+    # RuntimeError: Fail to enable Kineto Profiler on XPU due to error code: 200
+    # https://github.com/intel/torch-xpu-ops/issues/731
+    "test_profiler",
+    "test_record_function",
 )
 res += launch_test("test_autograd_xpu.py", skip_list)
 
@@ -1862,6 +1868,12 @@ def launch_test(test_case, skip_list=None, exe_list=None):
     # XPU does not support tunable.
     "test_bmm_tunableop_rocm_xpu_float32",
     "test_numeric_check_leak_tunableop_rocm_xpu_float32",
+
+    # CUDA bias cases added in latest PyTorch
+    # AttributeError: module 'torch._C' has no attribute '_cuda_tunableop_enable'
+    "test_matmul_check_entries_tunableop_xpu_float16",
+    "test_minimum_tuning_iteration_tunableop_xpu_float16",
+    "test_validator_tunableop_rocm_xpu_float32",
 )
 res += launch_test("test_linalg_xpu.py", skip_list)