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

[BugFix] Add async communication check for capturing mode (#8149)

Browse Source 
### What this PR does / why we need it?
Introduce a check to not using asynchronous communication under
`enable_dsa_cp_with_layer_shard` branch on capturing mode. This change
prevents potential stream and event issues when operating in
graph/capturing mode, ensuring safer communication practices.

### Does this PR introduce _any_ user-facing change?
No.

### How was this patch tested?
E2E test with dsv32 + FC1 + FULL_DECODE_ONLY +
kv_transfer_config(kv_both)

---------

Signed-off-by: chenchuw886 <chenchuw@huawei.com>
Co-authored-by: chenchuw886 <chenchuw@huawei.com>
This commit is contained in:
Frank Chen
2026-04-12 21:52:54 +08:00
committed by GitHub
parent c1f323ee46
commit 31186a3a9d
8 changed files with 85 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
tests/ut/test_platform.py
View File

@@ -438,6 +438,36 @@ class TestNPUPlatform(TestBase):
self.assertEqual(vllm_config.cache_config.block_size, 512)
def test_validate_layer_sharding_config_accepts_single_node(self):
vllm_config = TestNPUPlatform.mock_vllm_config()
vllm_config.additional_config = {"layer_sharding": ["q_b_proj", "o_proj"]}
vllm_config.kv_transfer_config = None
self.platform._validate_layer_sharding_config(vllm_config)
def test_validate_layer_sharding_config_accepts_kv_producer(self):
vllm_config = TestNPUPlatform.mock_vllm_config()
vllm_config.additional_config = {"layer_sharding": ["q_b_proj", "o_proj"]}
vllm_config.kv_transfer_config = MagicMock(is_kv_producer=True, kv_role="kv_producer")
self.platform._validate_layer_sharding_config(vllm_config)
def test_validate_layer_sharding_config_rejects_non_kv_producer(self):
vllm_config = TestNPUPlatform.mock_vllm_config()
vllm_config.additional_config = {"layer_sharding": ["q_b_proj", "o_proj"]}
vllm_config.kv_transfer_config = MagicMock(is_kv_producer=False, kv_role="kv_consumer")
with pytest.raises(ValueError, match="layer_sharding is only supported on P nodes"):
self.platform._validate_layer_sharding_config(vllm_config)
def test_validate_layer_sharding_config_rejects_kv_both(self):
vllm_config = TestNPUPlatform.mock_vllm_config()
vllm_config.additional_config = {"layer_sharding": ["q_b_proj", "o_proj"]}
vllm_config.kv_transfer_config = MagicMock(is_kv_producer=True, kv_role="kv_both")
with pytest.raises(ValueError, match="layer_sharding is only supported on P nodes"):
self.platform._validate_layer_sharding_config(vllm_config)
@patch("vllm_ascend.quantization.utils.maybe_auto_detect_quantization")
@patch("vllm_ascend.utils.get_ascend_device_type", return_value=AscendDeviceType.A3)
@patch("vllm_ascend.ascend_config.init_ascend_config")