EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Feat][SP] Suport SP for VL MoE models (#7044)

Browse Source 
### What this PR does / why we need it?

2nd PR for https://github.com/vllm-project/vllm-ascend/issues/5712,
extend SP to VL MoE models.


### Does this PR introduce _any_ user-facing change?
remove `sp_threshold` in additional config and reuse `sp_min_token_num`
from vLLM.


### How was this patch tested?
- Model: Qwen3-VL-30B-A3B, 
- TP4 DP2
- 100 reqs
- max concurrency 1

| Seq length | Mean TTFT (ms) main | Mean TTFT (ms) this PR |
|------------|---------------------|------------------------|
| 4k         | 429.40               | 323.3                  |
| 16k        | 1297.01              | 911.74                |

- vLLM version: v0.16.0
- vLLM main:
4034c3d32e

---------

Signed-off-by: realliujiaxu <realliujiaxu@163.com>
This commit is contained in:
realliujiaxu
2026-03-24 17:16:00 +08:00
committed by GitHub
parent 9615bc33fd
commit 5d12446573
21 changed files with 947 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
vllm_ascend/utils.py
View File

@@ -764,8 +764,8 @@ def matmul_allreduce_enable() -> bool:
return envs_ascend.VLLM_ASCEND_ENABLE_MATMUL_ALLREDUCE
def enable_sp_by_pass(vllm_config: VllmConfig):
return not vllm_config.model_config.enforce_eager and vllm_config.compilation_config.pass_config.enable_sp
def enable_sp_by_pass():
return get_ascend_config().enable_sp_by_pass
def enable_sp(vllm_config=None, enable_shared_expert_dp: bool = False) -> bool:
@@ -791,7 +791,7 @@ def enable_sp(vllm_config=None, enable_shared_expert_dp: bool = False) -> bool:
# TODO remove it after vllm has this func
def shared_expert_dp_enabled() -> bool:
return get_ascend_config().enable_shared_expert_dp or enable_sp()
return get_ascend_config().enable_shared_expert_dp or enable_sp() or enable_sp_by_pass()
def prefill_context_parallel_enable() -> bool: