[Ops][Refactor] Remove custom rotary_embedding operator

[wangxiyuan]

What this PR does / why we need it?

This PR removes the custom rotary_embedding operator and its associated C++ kernel implementation, PyTorch bindings, and tests.

The codebase now falls back to using the native torch_npu._npu_rotary_embedding implementation. This change simplifies the codebase by removing custom, platform-specific kernel code and relying on the standard NPU library implementation, which is presumably more optimized and easier to maintain.

Does this PR introduce any user-facing change?

No. This is an internal refactoring and does not introduce any user-facing changes.

How was this patch tested?

The tests for the custom rotary_embedding operator have been removed along with the operator itself. The correctness of the fallback to the native torch_npu implementation is verified by existing CI tests for attention layers and models that use rotary embeddings.

• vLLM version: v0.15.0
• vLLM main: https://github.com/vllm-project/vllm/commit/v0.15.0

[github-actions]

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

• A PR should do only one thing, smaller PRs enable faster reviews.
• Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
• Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.

[gemini-code-assist]

Summary of Changes

Hello @wangxiyuan, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request streamlines the codebase by completely removing a custom Rotary Embedding operation. This includes its C++ kernel implementation, associated C++ and Python bindings, meta-registrations, and all dedicated test files. The removal simplifies the project structure and maintenance, indicating a shift towards relying on alternative or upstream implementations for this functionality.

Highlights

• Custom Rotary Embedding Kernel Removal: The entire custom C++ kernel implementation for Rotary Embedding (pos_encoding_kernels.cpp) has been removed from the codebase.
• C++ Bindings and Registrations Clean-up: All declarations and implementations related to the custom Rotary Embedding operation have been removed from csrc/ops.h, csrc/torch_binding.cpp, and csrc/torch_binding_meta.cpp, including its registration with the PyTorch operator library.
• Test File Deletion: Both unit tests (tests/ut/ops/test_rotary_embedding.py) and end-to-end tests (tests/e2e/nightly/single_node/ops/singlecard_ops/test_rotary_embedding.py) specifically designed for the custom Rotary Embedding have been removed.
• Python Logic Simplification: The Python-side logic in vllm_ascend/ops/rotary_embedding.py that conditionally invoked the custom kernel has been removed, along with the _custom_rotary_embedding_enabled helper function.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog

• csrc/kernels/pos_encoding_kernels.cpp
  • Removed the entire file, which contained the custom Rotary Embedding kernel.
• csrc/ops.h
  • Removed the external declaration for rotary_embedding_impl.
• csrc/torch_binding.cpp
  • Removed the rotary_embedding function and its corresponding torch.ops._C_ascend registration.
• csrc/torch_binding_meta.cpp
  • Removed the rotary_embedding_meta function and its meta-implementation registration.
• tests/e2e/nightly/single_node/ops/singlecard_ops/test_rotary_embedding.py
  • Removed the end-to-end test file for the custom rotary embedding.
• tests/ut/ops/test_rotary_embedding.py
  • Removed the unit test file for the custom rotary embedding.
• vllm_ascend/meta_registration.py
  • Removed the rotary_embedding_meta function and its registration from the meta-registration utility.
• vllm_ascend/ops/rotary_embedding.py
  • Removed the _custom_rotary_embedding_enabled function.
  • Removed the conditional logic that called the custom torch.ops._C_ascend.rotary_embedding kernel.

Activity

• The pull request was opened by wangxiyuan with a title indicating a miscellaneous change to remove the custom Rotary Embedding operation.
• The description body is empty, suggesting the change is self-explanatory given the title and file modifications.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request cleanly removes the custom rotary_embedding operator, including its C++ kernel, PyTorch bindings, and associated tests. The implementation now falls back to the native torch_npu._npu_rotary_embedding operator. This is a good simplification of the codebase. The removal appears to be complete, with no dangling references left behind.

To improve the pull request, I recommend filling out the description to provide context for this change, as per the repository's contribution guidelines. I have provided a suggested title and summary below for your convenience.

Suggested PR Title:

[Ops][Refactor] Remove custom rotary_embedding operator

Suggested PR Summary:

### What this PR does / why we need it?
This PR removes the custom `rotary_embedding` operator and its associated C++ kernel implementation, PyTorch bindings, and tests.

The codebase now falls back to using the native `torch_npu._npu_rotary_embedding` implementation. This change simplifies the codebase by removing custom, platform-specific kernel code and relying on the standard NPU library implementation, which is presumably more optimized and easier to maintain.

### Does this PR introduce _any_ user-facing change?
No. This is an internal refactoring and does not introduce any user-facing changes.

### How was this patch tested?
The tests for the custom `rotary_embedding` operator have been removed along with the operator itself. The correctness of the fallback to the native `torch_npu` implementation is verified by existing CI tests for attention layers and models that use rotary embeddings.

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.