[Fix] Remove unnecessary NPU synchronization in MTP proposer #4325
+0
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…formances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations.. A more proper one is already implemented in vllm-project#4233. Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
Collaborator
Author
|
This is a follow up of former discussion. |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request proposes removing a torch.npu.synchronize() call within the speculative decoding loop in MtpProposer. This change aims to enhance performance by allowing greater overlap between CPU-based metadata preparation and NPU computation. Based on my analysis, the explicit synchronization is redundant because data dependencies are managed by the NPU stream, and a more suitable synchronization point is already in place for graph-based execution. Therefore, this modification is a sound performance optimization, and I have not identified any potential correctness issues.
|
👋 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. |
Copilot finished reviewing on behalf of
yiz-liu
November 21, 2025 04:18
There was a problem hiding this comment.
Pull Request Overview
This PR removes an unnecessary NPU synchronization call from the MTP proposer to improve pipeline efficiency by enabling better overlap between CPU and NPU operations. The PR description indicates that a more appropriate synchronization mechanism is already implemented in PR #4233.
- Removed
torch.npu.synchronize()call from the metadata preparation loop in the_proposemethod
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Collaborator
Author
|
There was an irrlevant lint error. |
NSDie
pushed a commit
to NSDie/vllm-ascend
that referenced
this pull request
Nov 24, 2025
…oject#4325) ### What this PR does / why we need it? Remove unnecessary NPU synchronization in MTP proposer to improve performances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in vllm-project#4233 ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None. - vLLM version: v0.11.0 - vLLM main: vllm-project/vllm@2918c1b Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com> Signed-off-by: nsdie <yeyifan@huawei.com>
NSDie
pushed a commit
to NSDie/vllm-ascend
that referenced
this pull request
Nov 24, 2025
…oject#4325) ### What this PR does / why we need it? Remove unnecessary NPU synchronization in MTP proposer to improve performances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in vllm-project#4233 ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None. - vLLM version: v0.11.0 - vLLM main: vllm-project/vllm@2918c1b Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com> Signed-off-by: nsdie <yeyifan@huawei.com>
NSDie
pushed a commit
to NSDie/vllm-ascend
that referenced
this pull request
Nov 24, 2025
…oject#4325) ### What this PR does / why we need it? Remove unnecessary NPU synchronization in MTP proposer to improve performances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in vllm-project#4233 ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None. - vLLM version: v0.11.0 - vLLM main: vllm-project/vllm@2918c1b Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com> Signed-off-by: nsdie <yeyifan@huawei.com>
NSDie
pushed a commit
to NSDie/vllm-ascend
that referenced
this pull request
Nov 24, 2025
…oject#4325) ### What this PR does / why we need it? Remove unnecessary NPU synchronization in MTP proposer to improve performances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in vllm-project#4233 ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None. - vLLM version: v0.11.0 - vLLM main: vllm-project/vllm@2918c1b Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com> Signed-off-by: nsdie <yeyifan@huawei.com>
NSDie
pushed a commit
to NSDie/vllm-ascend
that referenced
this pull request
Nov 24, 2025
…oject#4325) ### What this PR does / why we need it? Remove unnecessary NPU synchronization in MTP proposer to improve performances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in vllm-project#4233 ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None. - vLLM version: v0.11.0 - vLLM main: vllm-project/vllm@2918c1b Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com> Signed-off-by: nsdie <yeyifan@huawei.com>
845473182
pushed a commit
to 845473182/vllm-ascend
that referenced
this pull request
Nov 29, 2025
…oject#4325) ### What this PR does / why we need it? Remove unnecessary NPU synchronization in MTP proposer to improve performances. Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in vllm-project#4233 ### Does this PR introduce _any_ user-facing change? None. ### How was this patch tested? None. - vLLM version: v0.11.0 - vLLM main: vllm-project/vllm@2918c1b Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What this PR does / why we need it?
Remove unnecessary NPU synchronization in MTP proposer to improve performances.
Removing this synchronization point improves pipeline efficiency by allowing for better overlap between CPU and NPU operations. A more proper one is already implemented in #4233
Does this PR introduce any user-facing change?
None.
How was this patch tested?
None.