[EPLB]Eplb Config Renaming

[shenchuxiaofugui]

What this PR does / why we need it?

1. Rename num_iterations_eplb_update to expert_heat_collection_interval.
2. Rename num_wait_worker_iterations to algorithm_execution_interval.
3. Rename init_redundancy_expert to num_redundant_experts because the variable with the same meaning in vLLM is named this way.
4. Delete gate_eplb because we don't need this feature.
5. Move eplb config into a dict in additional config.
6. Depend on pr5817

Does this PR introduce any user-facing change?

before this pr：
--additional-config '{"dynamic_eplb":true, "num_iterations_eplb_update": 4000, "num_wait_worker_iterations": 150, "init_redundancy_expert": 16, "expert_map_path": "xxx.json"}'

after this pr:
--additional-config '{"eplb_config":{"dynamic_eplb":true,"expert_heat_collection_interval":4000, "algorithm_execution_interval":150,"num_redundant_experts": 16, "expert_map_path": "xxx.json"}}'

How was this patch tested?

test qwen3-235b eplb num_redundant_experts=16

without pr5817

dataset	version	metric	mode	vllm-api-general-chat
aime2024	604a78	accuracy	gen	83.33

with pr5817

dataset	version	metric	mode	vllm-api-general-chat
aime2024	604a78	accuracy	gen	86.67

• vLLM version: v0.13.0
• vLLM main: vllm-project/vllm@45c1ca1

[gemini-code-assist]

Code Review

This pull request refactors several configuration parameters for the Expert Parallelism Load Balancer (EPLB) to improve clarity. The changes are applied consistently across documentation, tests, and source code. My review focuses on ensuring the documentation is clear and accurate after these changes. I found some issues in the additional_config.md file, including duplicate entries and formatting errors, which could confuse users.

[gemini-code-assist]

The configuration table in this section has several issues that could be confusing for users:

• Duplicate Configurations:
  • enable_async_exponential is listed on line 35 (as a bool) and again on line 46 (as an int).
  • dump_config (line 45) appears to be a duplicate of dump_config_path (line 34), as they have identical descriptions. The correct parameter used in the code is dump_config_path.
  • lmhead_tensor_parallel_size (line 37) and oproj_tensor_parallel_size (line 38) are already documented under the finegrained_tp_config section below.
• Formatting and Typos:
  • The table formatting is misaligned from line 41 onwards, making it difficult to read.
  • There is a typo in the description for algorithm_execution_interval on line 42 ("The forward iterations...").

To improve clarity and correctness, I suggest replacing lines 37-46 to only include the relevant, correctly formatted, and de-duplicated configurations related to this PR's scope.

Suggested change

	| `lmhead_tensor_parallel_size` | int | `None` | The custom tensor parallel size of lmhead. Restriction: Can only be used when tensor_parallel=1 |
	| `oproj_tensor_parallel_size` | int | `None` | The custom tensor parallel size of oproj. |
	| `multistream_overlap_shared_expert` | bool | `False` | Whether to enable multistream shared expert. This option only takes effect on MoE models with shared experts. |
	| `dynamic_eplb` | bool | `False` | Whether to enable dynamic EPLB. |
	| `expert_heat_collection_interval` | int | `400` | Forward iterations when EPLB begins. |
	| `algorithm_execution_interval` | int | `30` | The forward iterations when the EPLB worker will finish CPU tasks. In our test default value 30 can cover most cases. |
	| `expert_map_record_path` | str | `None` | Save the expert load calculation results to a new expert table in the specified directory. |
	| `num_redundant_experts` | int | `0` | Specify redundant experts during initialization. |
	| `dump_config` | str | `None` | Configuration file path for msprobe dump(eager mode). |
	| `enable_async_exponential` | int | `0` | Whether to enable async exponential overlap. To enable async exponential, set this config to 1. |
	| `multistream_overlap_shared_expert` | bool | `False` | Whether to enable multistream shared expert. This option only takes effect on MoE models with shared experts. |
	| `dynamic_eplb` | bool | `False` | Whether to enable dynamic EPLB. |
	| `expert_heat_collection_interval` | int | `400` | Forward iterations when EPLB begins. |
	| `algorithm_execution_interval` | int | `30` | The forward iterations when the EPLB worker will finish CPU tasks. In our test default value 30 can cover most cases. |
	| `expert_map_record_path` | str | `None` | Save the expert load calculation results to a new expert table in the specified directory. |
	| `num_redundant_experts` | int | `0` | Specify redundant experts during initialization. |

[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.

[github-actions]

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

[github-actions]

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

[wangxiyuan]

ci passed here: https://github.com/vllm-project/vllm-ascend/actions/runs/20952178776?pr=5533
please rebase and fix the merge conflict. I'll merge it without CI test again.

[shenchuxiaofugui]

ci passed here: https://github.com/vllm-project/vllm-ascend/actions/runs/20952178776?pr=5533 please rebase and fix the merge conflict. I'll merge it without CI test again.

Conflict was resolved.