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

[Feature] support aclgraph for model runner v2 (#7110)

Browse Source 
### What this PR does / why we need it?
This PR aims to support aclgraph for model runner v2, please see RFC
#5208. The PR contains these modifications:
- adapt to newest commit of vllm main branch.
- supply a unified interface of extra forward context for both model
runner v1 and model runner v2.
- implement graph mode for main model. 

### Does this PR introduce _any_ user-facing change?
no

### How was this patch tested?

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

---------

Signed-off-by: Ronald1995 <ronaldautomobile@163.com>
This commit is contained in:
Ronald
2026-03-13 09:11:46 +08:00
committed by GitHub
parent 1f71da80eb
commit c980e68d40
52 changed files with 840 additions and 309 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
tests/ut/ops/test_fused_moe.py
View File

@@ -119,11 +119,9 @@ def mock_dist_env(mocker: MockerFixture):
return_value=(torch.tensor([0, 1, 2, -1, -1, -1, -1, -1]), None, 0)), \
patch('vllm_ascend.ops.fused_moe.fused_moe.get_forward_context',
return_value=mock_forward_context_obj), \
patch('vllm_ascend.ops.fused_moe.prepare_finalize.get_forward_context',
patch('vllm_ascend.ascend_forward_context.get_forward_context',
return_value=mock_forward_context_obj), \
patch("vllm_ascend.utils.get_ascend_device_type", return_value=AscendDeviceType.A3), \
patch('vllm_ascend.ops.fused_moe.moe_mlp.get_forward_context',
return_value=mock_forward_context_obj), \
patch('vllm_ascend.ops.fused_moe.moe_comm_method.MC2CommImpl._get_token_dispatcher',
return_value=None), \
patch('vllm_ascend.ops.fused_moe.moe_comm_method.AlltoAllCommImpl._get_token_dispatcher',
@@ -298,7 +296,7 @@ class TestUnifiedApplyMLP(TestBase):
@patch('vllm_ascend.ops.fused_moe.moe_mlp.get_weight_prefetch_method',
return_value=MagicMock())
@patch('vllm_ascend.ops.fused_moe.moe_mlp.get_forward_context')
@patch('vllm_ascend.ascend_forward_context.get_forward_context')
@patch('vllm_ascend.utils.get_ascend_device_type',
return_value=AscendDeviceType.A3)
@patch('torch_npu.npu_grouped_matmul')
@@ -407,7 +405,7 @@ class TestUnifiedApplyMLP(TestBase):
@patch('vllm_ascend.ops.fused_moe.moe_mlp.HAS_TRITON', False)
@patch('vllm_ascend.ops.fused_moe.moe_mlp.get_weight_prefetch_method',
return_value=MagicMock())
@patch('vllm_ascend.ops.fused_moe.moe_mlp.get_forward_context')
@patch('vllm_ascend.ascend_forward_context.get_forward_context')
@patch('torch_npu.npu_grouped_matmul')
@patch('torch_npu.npu_swiglu')
@patch('torch_npu.npu_dynamic_quant')
@@ -513,7 +511,7 @@ class TestUnifiedApplyMLP(TestBase):
@patch("vllm_ascend.ops.fused_moe.moe_mlp.get_weight_prefetch_method",
return_value=MagicMock())
@patch("vllm_ascend.ops.fused_moe.moe_mlp.get_forward_context")
@patch("vllm_ascend.ascend_forward_context.get_forward_context")
@patch("torch_npu.npu_grouped_matmul")
@patch("torch_npu.npu_swiglu")
@patch("torch_npu.npu_grouped_matmul_swiglu_quant")