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

[Misc] Clean up useless patch (#3320)

Browse Source 
### What this PR does / why we need it?
1. clean up v0.10.2 support in ut and e2e test
2. remove v0.11.0 period job, we're at v0.11.0 now.
3. remove uesless patch for deepseek v3.2. They have been done in vLLM
already.
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?


- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
wangxiyuan
2025-10-09 14:07:26 +08:00
committed by GitHub
parent a43e2f61e1
commit 1c5b302f0d
10 changed files with 29 additions and 412 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
tests/ut/ops/test_fused_moe_prepare_and_finalize.py
View File

@@ -8,7 +8,6 @@ from vllm_ascend.ops.moe.fused_moe_prepare_and_finalize import (
FusedMoEPrepareAndFinalizeWithAll2All,
FusedMoEPrepareAndFinalizeWithAllGather, FusedMoEPrepareAndFinalizeWithMC2,
FusedMoEPrepareAndFinalizeWithNaiveMulticast)
from vllm_ascend.utils import vllm_version_is
class TestFusedMoEPrepareAndFinalize(unittest.TestCase):
@@ -231,12 +230,8 @@ class TestFusedMoEPrepareAndFinalize(unittest.TestCase):
mock_get_dp_group):
# Mock forward context with DP metadata
mock_context = MagicMock()
if vllm_version_is("0.10.2"):
mock_context.dp_metadata.cu_tokens_across_dp_cpu = torch.tensor(
[2, 5, 7])
else:
mock_context.dp_metadata.cu_tokens_across_sp.return_value = torch.tensor(
[2, 5, 7])
mock_context.dp_metadata.cu_tokens_across_sp.return_value = torch.tensor(
[2, 5, 7])
mock_get_forward_context.return_value = mock_context
# Setup DP group mock

8
tests/ut/ops/test_fused_ops.py
View File

@@ -28,7 +28,7 @@ from vllm_ascend.ops.fused_moe import (AscendFusedMoE,
AscendUnquantizedFusedMoEMethod)
from vllm_ascend.ops.moe.experts_selector import select_experts
from vllm_ascend.ops.moe.moe_mlp import cumsum_group_list, unified_apply_mlp
from vllm_ascend.utils import AscendSocVersion, adapt_patch, vllm_version_is
from vllm_ascend.utils import AscendSocVersion, adapt_patch
adapt_patch(True)
@@ -92,11 +92,7 @@ def mock_dist_env(mocker: MockerFixture):
return hidden_states
mock_moe_comm_method.finalize.side_effect = mock_finalize
if vllm_version_is("0.10.2"):
dp_metadata = MagicMock(cu_tokens_across_dp_cpu=[5, 10])
else:
dp_metadata = MagicMock(num_tokens_across_dp_cpu=[5, 5])
dp_metadata = MagicMock(num_tokens_across_dp_cpu=[5, 5])
mock_forward_context_obj = MagicMock(moe_comm_method=mock_moe_comm_method,
moe_comm_type=MoECommType.MC2,
max_tokens_across_dp=10,