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

refactor: add a check before layer_sharding logging (#7186)

Browse Source 
### What this PR does / why we need it?
We should only display this log message when layer_sharding is enabled.
- vLLM version: v0.16.0
- vLLM main:
4034c3d32e

Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
This commit is contained in:
Qiu
2026-03-12 11:56:04 +08:00
committed by GitHub
parent 5f3826b093
commit aa0143e55d
1 changed files with 6 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
vllm_ascend/ascend_config.py
View File

@@ -52,11 +52,12 @@ class AscendConfig:
self.dump_config_path = additional_config.get("dump_config_path", None) self.dump_config_path = additional_config.get("dump_config_path", None)
self._construct_weight_prefetch_config(additional_config) self._construct_weight_prefetch_config(additional_config)
self.layer_sharding = additional_config.get("layer_sharding", None) self.layer_sharding = additional_config.get("layer_sharding", None)
logger.info_once( if self.layer_sharding:
f"Linear layer sharding enabled with config: {self.layer_sharding}. " logger.info_once(
"Note: This feature works optimally with FLASHCOMM2 and DSA-CP enabled; " f"Linear layer sharding enabled with config: {self.layer_sharding}. "
"using it without these features may result in significant performance degradation." "Note: This feature works optimally with FLASHCOMM2 and DSA-CP enabled; "
) "using it without these features may result in significant performance degradation."
)
self.enable_shared_expert_dp = ( self.enable_shared_expert_dp = (
additional_config.get("enable_shared_expert_dp", False) additional_config.get("enable_shared_expert_dp", False)