[tdalmia-aws]

Description

Marks the 7 TestTraceableCollectives Neuron tests that fail in the nightly distributed job as xfail.

Related Issue

Related to #1021

Motivation and Context

The threaded process group in these tests copies tensors between Neuron devices, and a cross-device copy currently keeps the result on the source device (#1021). The tests fail with Inconsistent device IDs found in kernel outputs: expected device 0 but found device 1, which keeps the nightly red: https://github.com/aws-neuron/torch-neuronx/actions/runs/41279502191

The copy fix will come with the next release. Until then these are xfail (not strict), so they report as xpass once the fix lands.

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)

Test-only change: skip registry entries.

Testing

Ran locally (no Neuron hardware needed):

pre-commit run --files tests/pytorch_tests_native/skip_registry/distributed_xfails_registry.py

All hooks pass.

Ran on Neuron hardware:

python -m pytest tests/pytorch_tests_native/test/distributed/test_functional_api.py -k TestTraceableCollectives

On main these 7 fail with the error above, the same as the nightly. The xfail entries cover exactly those 7 tests.

Checklist

• Added or updated tests for this change
• I've performed a self-review for my code
• Ran the tests, or explained under Testing why I could not
• Updated the corresponding documentation
• My changes generate no new warnings
• Said under Description if an AI assistant helped with this change (see AI_POLICY.md)