xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-18 20:57:19 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix for cpu

Browse Source 
Signed-off-by: baonudesifeizhai <baonudesifeizhai@gmail.com>
This commit is contained in:
baonudesifeizhai 2025-12-23 15:58:39 -05:00
parent a2774c4d69
commit ff82fce3b2
2 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
tests/compile/fullgraph/test_full_graph.py
View File

@ -13,9 +13,7 @@ from vllm import LLM, SamplingParams
from vllm.attention.backends.registry import AttentionBackendEnum
from vllm.config import CompilationConfig, CompilationMode, CUDAGraphMode, PassConfig
from vllm.platforms import current_platform
from vllm.utils.torch_utils import (
is_torch_equal_or_newer,
)
from vllm.utils.torch_utils import is_torch_equal_or_newer
from ...utils import create_new_process_for_each_test

7
vllm/env_override.py
View File

@ -377,9 +377,12 @@ def _patch_get_raw_stream_if_needed():
if is_torch_equal("2.9.0") or is_torch_equal("2.9.1"):
import builtins
from torch._C import _cuda_getCurrentRawStream as _get_raw_stream
# Check if CUDA functionality is available without initializing CUDA
# _cuda_getCurrentRawStream only exists in CUDA builds of PyTorch
if hasattr(torch._C, "_cuda_getCurrentRawStream"):
from torch._C import _cuda_getCurrentRawStream as _get_raw_stream
builtins.get_raw_stream = _get_raw_stream
builtins.get_raw_stream = _get_raw_stream
_patch_get_raw_stream_if_needed()