biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V0 deprecation] Deprecate V0 Neuron backend (#21159)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon
2025-09-06 16:15:18 -07:00
committed by GitHub
parent 848562bd49
commit 4172235ab7
46 changed files with 10 additions and 5462 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
setup.py
View File

@@ -413,8 +413,7 @@ def _no_device() -> bool:
def _is_cuda() -> bool:
has_cuda = torch.version.cuda is not None
return (VLLM_TARGET_DEVICE == "cuda" and has_cuda
and not (_is_neuron() or _is_tpu()))
return (VLLM_TARGET_DEVICE == "cuda" and has_cuda and not _is_tpu())
def _is_hip() -> bool:
@@ -422,10 +421,6 @@ def _is_hip() -> bool:
or VLLM_TARGET_DEVICE == "rocm") and torch.version.hip is not None
def _is_neuron() -> bool:
return VLLM_TARGET_DEVICE == "neuron"
def _is_tpu() -> bool:
return VLLM_TARGET_DEVICE == "tpu"
@@ -470,25 +465,6 @@ def get_rocm_version():
return None
def get_neuronxcc_version():
import sysconfig
site_dir = sysconfig.get_paths()["purelib"]
version_file = os.path.join(site_dir, "neuronxcc", "version",
"__init__.py")
# Check if the command was executed successfully
with open(version_file) as fp:
content = fp.read()
# Extract the version using a regular expression
match = re.search(r"__version__ = '(\S+)'", content)
if match:
# Return the version string
return match.group(1)
else:
raise RuntimeError("Could not find Neuron version in the output")
def get_nvcc_cuda_version() -> Version:
"""Get the CUDA version from nvcc.
@@ -541,12 +517,6 @@ def get_vllm_version() -> str:
rocm_version = get_rocm_version() or torch.version.hip
if rocm_version and rocm_version != MAIN_CUDA_VERSION:
version += f"{sep}rocm{rocm_version.replace('.', '')[:3]}"
elif _is_neuron():
# Get the Neuron version
neuron_version = str(get_neuronxcc_version())
if neuron_version != MAIN_CUDA_VERSION:
neuron_version_str = neuron_version.replace(".", "")[:3]
version += f"{sep}neuron{neuron_version_str}"
elif _is_tpu():
version += f"{sep}tpu"
elif _is_cpu():
@@ -591,8 +561,6 @@ def get_requirements() -> list[str]:
requirements = modified_requirements
elif _is_hip():
requirements = _read_requirements("rocm.txt")
elif _is_neuron():
requirements = _read_requirements("neuron.txt")
elif _is_tpu():
requirements = _read_requirements("tpu.txt")
elif _is_cpu():
@@ -601,7 +569,7 @@ def get_requirements() -> list[str]:
requirements = _read_requirements("xpu.txt")
else:
raise ValueError(
"Unsupported platform, please use CUDA, ROCm, Neuron, or CPU.")
"Unsupported platform, please use CUDA, ROCm, or CPU.")
return requirements