Delete deprecated Int4XPULayout and related code

[jerryzh168]

Stack from ghstack (oldest at bottom):

• Delete deprecated TensorCoreTiledLayout and related code #4153
• Move bitpacking.py to prototype and add uintx_utils.py #4152
• Delete deprecated PlainLayout, PlainAQTTensorImpl and related v1 code paths #4151
• Delete deprecated PackedLinearInt8DynamicActivationIntxWeightLayout and related code #4150
• Delete deprecated Int4CPULayout and related code #4149
• Delete deprecated SemiSparseLayout and related code #4148
• -> Delete deprecated Int4XPULayout and related code #4147

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4147

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[jerryzh168]

@claude review

[claude]

Claude finished @jerryzh168's task in 2m 21s —— View job

---

Review: Delete deprecated Int4XPULayout and related code

• Gather context and read the diff
• Check for remaining references to deleted code
• Review each changed file
• Post review summary

---

LGTM - This is a clean deletion of deprecated code with no loose ends.

Summary

This PR removes the deprecated Int4XPULayout class and all associated code (481 lines deleted). The changes span 5 files and are straightforward.

File-by-file review

1. torchao/dtypes/uintx/int4_xpu_layout.py (deleted) - The entire file is removed. It contained Int4XPULayout, Int4XPUAQTTensorImpl, and two linear dispatch implementations. All of these were already marked deprecated with warnings pointing to issue #2752.

2. torchao/dtypes/uintx/__init__.py - Clean removal of the Int4XPULayout import and __all__ entry. Remaining exports are intact.

3. torchao/dtypes/__init__.py - Clean removal of Int4XPULayout from the import and __all__ list.

4. torchao/dtypes/affine_quantized_tensor_ops.py - Removes the import of the four XPU-specific dispatch functions and their registration entries in _register_aqt_quantized_linear_dispatches(). The remaining dispatch registrations are unaffected.

5. torchao/quantization/quant_api.py - Removes Int4XPULayout from the import, LAYOUT_TO_ZERO_POINT_DOMAIN, and LAYOUT_TO_PRESERVE_ZEROS dictionaries. Both dictionaries are still used by QAT code (torchao/quantization/qat/fake_quantize_config.py:393-397) with the remaining layouts.

Verification

I searched the entire codebase for any remaining references to Int4XPULayout, int4_xpu_layout, or Int4XPUAQTTensorImpl - no references remain. The deletion is complete.

No issues found.

---