DISABLED Test_var_mean_tile_reduction_True_dynamic_shapes_cuda (main.DynamicShapesGPUTests)

Mar 12, 2025 by ADMIN 96 views

Additional Information

This test was disabled due to its failure in Continuous Integration (CI) environments. To understand the current status of this test, please refer to the recent examples and workflow logs provided below.

Platforms

rocm: This test is specifically designed for the ROCm platform.

Recent Examples and Workflow Logs

Recent Examples: https://hud.pytorch.org/flakytest?name=test_var_mean_tile_reduction_True_dynamic_shapes_cuda&suite=DynamicShapesGPUTests&limit=100
Workflow Logs: https://github.com/pytorch/pytorch/runs/38592796229

Debugging Instructions

To debug this test, follow these steps:

Click on the recent samples link: This will take you to the recent examples of this test.
Click on the workflow logs linked above: This will take you to the workflow logs of this test.
Click on the Test step of the job so that it is expanded: This will allow you to grep for the relevant log snippets.
Grep for test_var_mean_tile_reduction_True_dynamic_shapes_cuda: This will show you the log snippets related to this test.
Study the logs: Analyze the logs to understand the failure of this test.

Sample Error Message

Traceback (most recent call last):
  File "/var/lib/jenkins/pytorch/test/inductor/test_torchinductor.py", line 5370, in test_var_mean
    self.common(
  File "/opt/conda/envs/py_3.10/lib/python3.10/contextlib.py", line 79, in inner
    return func(*args, **kwds)
  File "/var/lib/jenkins/pytorch/test/inductor/test_torchinductor.py", line 631, in check_model_gpu
    check_model(
  File "/var/lib/jenkins/pytorch/test/inductor/test_torchinductor.py", line 513, in check_model
    self.assertEqual(
  File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/testing/_internal/common_utils.py", line 4091, in assertEqual
    raise error_metas.pop()[0].to_error(  # type: ignore[index]
AssertionError: Tensor-likes are not close!

Mismatched elements: 4 / 4 (100.0%)
Greatest absolute difference: 0.6583480834960938 at index (0, 2) (up to 1e-05 allowed)
Greatest relative difference: 0.8614732623100281 at index (0, 0) (up to 1.3e-06 allowed)

The failure occurred for item [2]

To execute this test, run the following from the base repo dir:
    PYTORCH_TEST_WITH_ROCM=1 python test/inductor/test_torchinductor_dynamic_shapes.py DynamicShapesGPUTests.test_var_mean_tile_reduction_True_dynamic_shapes_cuda

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

Test File Path

inductor/test_torchinductor_dynamic_shapes.py: This is the test file path for this test.

CC

@jeffdaily: This test is being CC'd to Jeff Daily.
@sunway513: This test is being CC'd to Sunway513.
@jithunnair-amd: This test is being CC'd to Jithunnair-amd.
@pruthvistony: This test is being CC'd to Pruthvistony.
@ROCmSupport: This test is being CC'd to ROCmSupport.
@dllehr-amd: This test is being CC'd to Dllehr-amd.
@jataylo: This test is being CC'd to Jataylo.
@hongxiayang: This test is being CC'd to Hongxiayang.
@naromero77amd: This test is being CC'd to Naromero77amd.
@clee2000: This test is being CC'd to Clee2000.
@voznesenskym: This test is being CC'd to Voznesenskym.
@penguinwu: This test is being CC'd to Penguinwu.
@EikanWang: This test is being CC'd to EikanWang.
@jgong5: This test is being CC'd to Jgong5.
@Guobing-Chen: This test is being CC'd to Guobing-Chen.
@XiaobingSuper: This test is being CC'd to XiaobingSuper.
@zhuhaozhe: This test is being CC'd to Zhuhaozhe.
@blzheng: This test is being CC'd to Blzheng.
@wenzhe-nrv: This test is being CC'd to Wenzhe-nrv.
@jiayisunx: This test is being CC'd to Jiayisunx.
@ipiszy: This test is being CC'd to Ipiszy.
@chenyang78: This test is being CC'd to Chenyang78.
@kadeng: This test is being CC'd to Kadeng.
@muchulee8: This test is being CC'd to Muchulee8.
@amjames: This test is being CC'd to Amjames.
@chauhang: This test is being CC'd to Chauhang.
@aakhundov: This test is being CC'd to Aakhundov.

Q: What is the current status of the test_var_mean_tile_reduction_True_dynamic_shapes_cuda test?

A: The test is currently disabled due to its failure in Continuous Integration (CI) environments.

Q: Why is the test disabled?

A: The test is disabled because it is failing in CI environments. This means that the test is not passing the required tests and is causing issues in the CI pipeline.

Q: What are the recent examples of this test?

A: The recent examples of this test can be found at https://hud.pytorch.org/flakytest?name=test_var_mean_tile_reduction_True_dynamic_shapes_cuda&suite=DynamicShapesGPUTests&limit=100.

Q: What are the workflow logs for this test?

A: The workflow logs for this test can be found at https://github.com/pytorch/pytorch/runs/38592796229.

Q: How can I debug this test?

A: To debug this test, follow these steps:

Click on the recent samples link to view the recent examples of this test.
Click on the workflow logs linked above to view the workflow logs of this test.
Click on the Test step of the job so that it is expanded to view the relevant log snippets.
Grep for test_var_mean_tile_reduction_True_dynamic_shapes_cuda to view the log snippets related to this test.
Study the logs to understand the failure of this test.

Q: What is the sample error message for this test?

A: The sample error message for this test is as follows:

Traceback (most recent call last):
  File "/var/lib/jenkins/pytorch/test/inductor/test_torchinductor.py", line 5370, in test_var_mean
    self.common(
  File "/opt/conda/envs/py_3.10/lib/python3.10/contextlib.py", line 79, in inner
    return func(*args, **kwds)
  File "/var/lib/jenkins/pytorch/test/inductor/test_torchinductor.py", line 631, in check_model_gpu
    check_model(
  File "/var/lib/jenkins/pytorch/test/inductor/test_torchinductor.py", line 513, in check_model
    self.assertEqual(
  File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/testing/_internal/common_utils.py", line 4091, in assertEqual
    raise error_metas.pop()[0].to_error(  # type: ignore[index]
AssertionError: Tensor-likes are not close!

Mismatched elements: 4 / 4 (100.0%)
Greatest absolute difference: 0.6583480834960938 at index (0, 2) (up to 1e-05 allowed)
Greatest relative difference: 0.8614732623100281 at index (0, 0) (up to 1.3e-06 allowed)

The failure occurred for item [2]

To execute this test, run the following from the base repo dir:
    PYTORCH_TEST_WITH_ROCM=1 python test/inductor/test_torchinductor_dynamic_shapes.py DynamicShapesGPUTests.test_var_mean_tile_reduction_True_dynamic_shapes_cuda

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

Q: What is the test file path for this test?

A: The test file path for this test is inductor/test_torchinductor_dynamic_shapes.py.

Q: Who is being CC'd for this test?

A: The following individuals are being CC'd for this test:

@jeffdaily: Jeff Daily
@sunway513: Sunway513
@jithunnair-amd: Jithunnair-amd
@pruthvistony: Pruthvistony
@ROCmSupport: ROCmSupport
@dllehr-amd: Dllehr-amd
@jataylo: Jataylo
@hongxiayang: Hongxiayang
@naromero77amd: Naromero77amd
@clee2000: Clee2000
@voznesenskym: Voznesenskym
@penguinwu: Penguinwu
@EikanWang: EikanWang
@jgong5: Jgong5
@Guobing-Chen: Guobing-Chen
@XiaobingSuper: XiaobingSuper
@zhuhaozhe: Zhuhaozhe
@blzheng: Blzheng
@wenzhe-nrv: Wenzhe-nrv
@jiayisunx: Jiayisunx
@ipiszy: Ipiszy
@chenyang78: Chenyang78
@kadeng: Kadeng
@muchulee8: Muchulee8
@amjames: Amjames
@chauhang: Chauhang
@aakhundov: Aakhundov