xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-15 04:15:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[XPU] Conditionally import CUDA-specific passes to avoid import errors on xpu platform (#21036)

Browse Source 
Signed-off-by: chzhang <chaojun.zhang@intel.com>
This commit is contained in:
Chaojun Zhang 2025-07-24 23:23:36 +08:00 committed by GitHub
parent 1b25f1fe75
commit d9f9a3fd96
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 6 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
vllm/compilation/pass_manager.py
View File

@ -5,12 +5,15 @@ from torch import fx as fx
from vllm.config import VllmConfig from vllm.config import VllmConfig
from vllm.logger import init_logger from vllm.logger import init_logger
from vllm.platforms import current_platform
if current_platform.is_cuda_alike():
from .fusion import FusionPass
from .collective_fusion import AllReduceFusionPass, AsyncTPPass
from .fusion_attn import AttnFusionPass
from .activation_quant_fusion import ActivationQuantFusionPass from .activation_quant_fusion import ActivationQuantFusionPass
from .collective_fusion import AllReduceFusionPass, AsyncTPPass
from .fix_functionalization import FixFunctionalizationPass from .fix_functionalization import FixFunctionalizationPass
from .fusion import FusionPass
from .fusion_attn import AttnFusionPass
from .inductor_pass import CustomGraphPass, InductorPass, get_pass_context from .inductor_pass import CustomGraphPass, InductorPass, get_pass_context
from .noop_elimination import NoOpEliminationPass from .noop_elimination import NoOpEliminationPass
from .sequence_parallelism import SequenceParallelismPass from .sequence_parallelism import SequenceParallelismPass