[BugFix]Revise the EPLB feature guide content and optimize the EPLB algorithm. #3080
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
added 30 commits
September 16, 2025 14:46
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
# Conflicts: # vllm_ascend/worker/model_runner_v1.py
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
ec when usingglobal_redundant_expert_num. Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Signed-off-by: offline0806 <z00858301@china.huawei.com>
# Conflicts: # docs/source/user_guide/configuration/additional_config.md # docs/source/user_guide/feature_guide/eplb_swift_balancer.md # vllm_ascend/eplb/core/eplb_utils.py # vllm_ascend/eplb/core/eplb_worker.py # vllm_ascend/ops/common_fused_moe.py # vllm_ascend/quantization/w8a8_dynamic.py
|
👋 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. |
github-actions
bot
added
documentation
![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 introduces several improvements to the EPLB (Expert Parallel Load Balancing) feature. It updates the documentation with new configuration parameters and provides clearer examples. A key change is the optimization of the generate_log2phy_map function, which is now vectorized for better performance. A correctness fix is also included in the EPLB worker. My review identifies a critical issue in one of the new documentation examples where the provided JSON is invalid, which could lead to user errors.
| vllm serve Qwen/Qwen3-235B-A22 \ | ||
| --tensor-parallel-size 16 \ | ||
| --enable-expert-parallel \ | ||
| --additional-config '{ "expert_map_record_path": "/path/to/eplb.json", "init_redundancy_expert": 16, dynamic_eplb":true,"num_iterations_eplb_update":400, "gate_eplb":true, "num_wait_worker_iterations":30}' |
There was a problem hiding this comment.
The JSON string in this example is invalid. There is a missing double quote " before dynamic_eplb. This will cause a parsing error for users who copy and paste this command.
Suggested change
| --additional-config '{ "expert_map_record_path": "/path/to/eplb.json", "init_redundancy_expert": 16, dynamic_eplb":true,"num_iterations_eplb_update":400, "gate_eplb":true, "num_wait_worker_iterations":30}' | |
| --additional-config '{ "expert_map_record_path": "/path/to/eplb.json", "init_redundancy_expert": 16, "dynamic_eplb":true,"num_iterations_eplb_update":400, "gate_eplb":true, "num_wait_worker_iterations":30}' |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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?
1.Revise the EPLB feature guide content.Add eplb params to ascend config.
2.Optimize the EPLB algorithm.
Does this PR introduce any user-facing change?
How was this patch tested?
We run vllm online serving with quantized qwen3_235b.
vLLM version: v0.10.2
vLLM main: vllm-project/vllm@c60e613