pytorch - 💡(How to fix) Fix DISABLED test_compile_with_user_stream_context (__main__.TestUserStreamCompile) [1 comments, 1 participants]

Traceback (most recent call last): File "/__w/pytorch/pytorch/test/inductor/test_user_streams.py", line 225, in test_compile_with_user_stream_context self.assertEqual(result, expected) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/test_case.py", line 121, in assertEqual return super().assertEqual(x, y, *args, **kwargs) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/testing/_internal/common_utils.py", line 4511, in assertEqual raise error_metas.pop()[0].to_error( # type: ignore[index] AssertionError: Tensor-likes are not close!

Mismatched elements: 1024 / 1024 (100.0%) Greatest absolute difference: 24.741540908813477 at index (404,) (up to 1e-05 allowed) Greatest relative difference: 1174.2518310546875 at index (802,) (up to 1.3e-06 allowed)

To execute this test, run the following from the base repo dir: python test/inductor/test_user_streams.py TestUserStreamCompile.test_compile_with_user_stream_context

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TL;DR

The test test_compile_with_user_stream_context is failing due to assertion errors, and the issue may be related to numerical instability or precision issues in the comparison of tensor-like objects.

Guidance

• Review the test case in inductor/test_user_streams.py and examine the comparison of result and expected in the test_compile_with_user_stream_context method to understand the source of the assertion error.
• Investigate the numerical instability or precision issues that may be causing the mismatch between the expected and actual values, potentially by increasing the allowed tolerance in the assertEqual method.
• Run the test locally using the provided command python test/inductor/test_user_streams.py TestUserStreamCompile.test_compile_with_user_stream_context to reproduce the error and gather more information.
• Check the workflow logs linked in the issue to see if there are any patterns or common factors in the failures.

Notes

The issue may be related to the specific environment or configuration used in the CI workflow, so it's essential to review the logs and test results carefully to identify any potential causes.

Recommendation

Apply workaround: Increase the allowed tolerance in the assertEqual method to account for potential numerical instability or precision issues, and re-run the test to see if the issue is resolved.