pytorch - 💡(How to fix) Fix DISABLED test_throw_on_cudamalloc_oom (__main__.TestCudaAllocator) [1 comments, 2 participants]

TL;DR

The test test_throw_on_cudamalloc_oom in test_cuda_expandable_segments.py is failing on the main branch and needs to be investigated and fixed.

Guidance

• Review the recent failure examples on torch-ci.com to understand the failure pattern and potential causes.
• Investigate the test_throw_on_cudamalloc_oom test case to identify the root cause of the failure, potentially related to CUDA memory allocation.
• Check the test environment and configuration to ensure it matches the expected setup for the test.
• Consider re-enabling the test and running it locally to reproduce and debug the issue.

Notes

The issue lacks specific technical details, and without more information, it's challenging to provide a precise fix or workaround.

Recommendation

Apply workaround: Temporarily disable the failing test until the root cause is identified and fixed, to prevent blocking the main branch.