Conversation
wangxiyuan
requested review from
realliujiaxu,
whx-sjtu and
zzzzwwjj
as code owners
Summary of ChangesHello @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 aims to fix a broken Continuous Integration (CI) pipeline by removing the registration of a specific custom operation. The change is focused on resolving an issue that was preventing CI from passing, ensuring the build process can complete successfully. Highlights
🧠 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
Activity
Using Gemini Code AssistThe 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
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 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
|
|
👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:
If CI fails, you can run linting and testing checks locally according Contributing and Testing. |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request aims to fix a broken CI by removing the registration of the prefetch_preprocess custom operator. However, this change introduces a critical issue. The operator is still called in other parts of the codebase (vllm_ascend/ops/weight_prefetch.py). If the code paths that use this operator are ever executed (e.g., by enabling weight prefetching), the application will crash due to the missing operator registration. This creates a latent bug. A more robust solution would be to either fix the operator's implementation or remove its usage entirely if the feature is being deprecated.
In line with the repository's style guide (lines 14-39 for summary, 41-61 for title), I am also suggesting an updated Pull Request title and summary to better describe the changes. The current PR description is empty, which violates the repository's contribution guidelines.
Suggested PR Title:
[CI][BugFix] Remove registration of prefetch_preprocess custom opSuggested PR Summary:
### What this PR does / why we need it?
This PR removes the registration of the `prefetch_preprocess` custom PyTorch operator. This is intended to resolve a CI failure caused by this operator.
By removing the registration, calls to `torch.ops.vllm.prefetch_preprocess` will be disabled, which is expected to make the CI pass.
### Does this PR introduce _any_ user-facing change?
No, assuming the weight prefetching feature was already broken or disabled. If a user tries to enable this feature, the application will now crash instead of failing silently or with a different error.
### How was this patch tested?
CI is expected to pass with this change.I am having trouble creating individual review comments. Click here to see my feedback.
vllm_ascend/ops/register_custom_ops.py (226-232)
Removing the registration for the prefetch_preprocess custom operator will cause runtime errors because the operator torch.ops.vllm.prefetch_preprocess is still being called in vllm_ascend/ops/weight_prefetch.py.
While this change might fix a CI issue by effectively disabling a broken feature, it introduces a latent bug. If weight prefetching is enabled, the application will crash because the custom op is not found.
The correct approach would be to either:
- Fix the implementation of
_prefetch_preprocess_implif it's broken. - If the feature is being deprecated, remove all calls to
torch.ops.vllm.prefetch_preprocessfromvllm_ascend/ops/weight_prefetch.pyto eliminate dead code.
Simply removing the registration is not a safe or complete solution.
…llm-project#6129)" This reverts commit 4fb3d5e.
…to qwen3next_rebase * 'main' of https://github.com/vllm-project/vllm-ascend: [Patch] Remove the patch of MiniCPM (vllm-project#5975) [P/D] layerwise connector support recompute scheduler (vllm-project#5900) [CI] Add workflow support for lint image build (vllm-project#6489) [Bugfix] Fix problematic dummy_run & improper input_batch_size in eagle (vllm-project#6517) [Refactor]310p_e2e test case update (vllm-project#6539) [Refactor]refactor p2p connector (vllm-project#6551) [Refactor]refactor 310p attention impl and add ut (vllm-project#6579) [Refactor]refactor 310p ops and add ut (vllm-project#6591) [Ops][Refactor] Remove custom rotary_embedding operator (vllm-project#6523) [Lint]Style: Convert `vllm-ascend/` to ruff format(new Batch vllm-project#8) (vllm-project#6604) [Test] Add initial multi modal cases of Qwen2.5-VL-7B-Instruct for disaggregated encoder (vllm-project#5301) [CI] Fix broken CI (vllm-project#6599) [Lint]Style: Convert `vllm-ascend/` to ruff format(Batch vllm-project#10) (vllm-project#6173) [Lint]Style: Convert `vllm-ascend/` to ruff format(Batch vllm-project#11) (vllm-project#6176) [Lint]Style: Convert `vllm-ascend/` to ruff format(Batch vllm-project#8) (vllm-project#6129) [Lint]Style: Convert `vllm-ascend/` to ruff format(Batch vllm-project#7) (vllm-project#6023) [CI][Misc] Some improvement for github action (vllm-project#6587) [Image] Bump mooncake version to v0.3.8.post1 (vllm-project#6428)
Revert vllm-project@4fb3d5e it breaks E2E Test - vLLM version: v0.15.0 - vLLM main: vllm-project/vllm@d7e17aa Signed-off-by: momochenchuw <chenchuw@huawei.com>
Revert vllm-project@4fb3d5e it breaks E2E Test - vLLM version: v0.15.0 - vLLM main: vllm-project/vllm@d7e17aa Signed-off-by: zrj026 <zhangrunjiang026@gmail.com>
Revert vllm-project@4fb3d5e it breaks E2E Test - vLLM version: v0.15.0 - vLLM main: vllm-project/vllm@d7e17aa
Revert vllm-project@4fb3d5e it breaks E2E Test - vLLM version: v0.15.0 - vLLM main: vllm-project/vllm@d7e17aa Signed-off-by: zrj026 <zhangrunjiang026@gmail.com>
Revert vllm-project@4fb3d5e it breaks E2E Test - vLLM version: v0.15.0 - vLLM main: vllm-project/vllm@d7e17aa
1 participant
Revert 4fb3d5e
it breaks E2E Test