[Bugfix][Intel] Fix XPU Dockerfile Build (#7824)

Signed-off-by: tylertitsworth <tyler.titsworth@intel.com>
Co-authored-by: youkaichao <youkaichao@126.com>

.buildkite/run-xpu-test.sh

@@ -11,4 +11,4 @@ trap remove_docker_container EXIT
 remove_docker_container
 
 # Run the image and launch offline inference
-docker run --network host --name xpu-test --device /dev/dri -v /dev/dri/by-path:/dev/dri/by-path xpu-test python3 examples/offline_inference.py
+docker run --network host --name xpu-test --device /dev/dri -v /dev/dri/by-path:/dev/dri/by-path --entrypoint="" xpu-test python3 examples/offline_inference.py

.dockerignore

-vllm/*.so
+/.github/
 /.venv
 /build
 dist
+Dockerfile*
+vllm/*.so

Dockerfile.xpu

-FROM intel/oneapi-basekit:2024.2.1-0-devel-ubuntu22.04
+FROM intel/oneapi-basekit:2024.2.1-0-devel-ubuntu22.04 AS vllm-base
 
 RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | tee /usr/share/keyrings/intel-oneapi-archive-keyring.gpg > /dev/null && \
     echo "deb [signed-by=/usr/share/keyrings/intel-oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main " | tee /etc/apt/sources.list.d/oneAPI.list && \
@@ -7,20 +7,49 @@ RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRO
     echo "deb [arch=amd64,i386 signed-by=/usr/share/keyrings/intel-graphics.gpg] https://repositories.intel.com/graphics/ubuntu jammy arc" | tee /etc/apt/sources.list.d/intel.gpu.jammy.list && \
     chmod 644 /usr/share/keyrings/intel-graphics.gpg
 
-RUN apt-get update  -y && \
-    apt-get install -y curl libicu70 lsb-release git wget vim numactl python3 python3-pip ffmpeg libsm6 libxext6 libgl1
-
-COPY ./ /workspace/vllm
+RUN apt-get update -y && \
+    apt-get install -y --no-install-recommends --fix-missing \
+    curl \
+    ffmpeg \
+    git \
+    libsndfile1 \
+    libsm6 \
+    libxext6 \
+    libgl1 \
+    lsb-release \
+    numactl \
+    python3 \
+    python3-dev \
+    python3-pip \
+    # vim \
+    wget
 
 WORKDIR /workspace/vllm
+COPY requirements-xpu.txt /workspace/vllm/requirements-xpu.txt
+COPY requirements-common.txt /workspace/vllm/requirements-common.txt
 
 RUN --mount=type=cache,target=/root/.cache/pip \
-    pip install -v --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/ \
-        cmake>=3.26 ninja packaging setuptools-scm>=8 wheel jinja2 \
-        -r requirements-xpu.txt
+    pip install --no-cache-dir \
+    --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/ \
+    -r requirements-xpu.txt
+
+COPY ./ /workspace/vllm
+
+ENV VLLM_TARGET_DEVICE=xpu
 
 RUN --mount=type=cache,target=/root/.cache/pip \
     --mount=type=bind,source=.git,target=.git \
-    VLLM_TARGET_DEVICE=xpu python3 setup.py install
+    python3 setup.py install
 
 CMD ["/bin/bash"]
+
+FROM vllm-base AS vllm-openai
+
+# install additional dependencies for openai api server
+RUN --mount=type=cache,target=/root/.cache/pip \
+    pip install accelerate hf_transfer 'modelscope!=1.15.0'
+
+ENV VLLM_USAGE_SOURCE production-docker-image \
+    TRITON_XPU_PROFILE 1
+
+ENTRYPOINT ["python3", "-m", "vllm.entrypoints.openai.api_server"]

requirements-common.txt

 psutil
 sentencepiece  # Required for LLaMA tokenizer.
 numpy < 2.0.0
-requests
+requests >= 2.26.0
 tqdm
 py-cpuinfo
 transformers >= 4.45.0  # Required for Llama 3.2.

requirements-xpu.txt

 # Common dependencies
 -r requirements-common.txt
 
-setuptools < 70.0.0 # IPEX's torch have some dependency. to be removed.
-
 ray >= 2.9
+cmake>=3.26
+ninja
+packaging
+setuptools-scm>=8
+wheel
+jinja2
 # Following pkgs retrieved from https://pytorch-extension.intel.com/release-whl/stable/xpu/us/
 torch == 2.3.1+cxx11.abi
 intel-extension-for-pytorch == 2.3.110+xpu

setup.py

@@ -415,6 +415,8 @@ def get_requirements() -> List[str]:
         for line in requirements:
             if line.startswith("-r "):
                 resolved_requirements += _read_requirements(line.split()[1])
+            elif line.startswith("--"):
+                continue
             else:
                 resolved_requirements.append(line)
         return resolved_requirements

vllm/platforms/__init__.py

@@ -42,6 +42,15 @@ try:
 except Exception:
     pass
 
+is_xpu = False
+
+try:
+    import torch
+    if hasattr(torch, 'xpu') and torch.xpu.is_available():
+        is_xpu = True
+except Exception:
+    pass
+
 is_cpu = False
 try:
     from importlib.metadata import version
@@ -60,6 +69,9 @@ elif is_cuda:
 elif is_rocm:
     from .rocm import RocmPlatform
     current_platform = RocmPlatform()
+elif is_xpu:
+    from .xpu import XPUPlatform
+    current_platform = XPUPlatform()
 elif is_cpu:
     from .cpu import CpuPlatform
     current_platform = CpuPlatform()

vllm/platforms/interface.py

@@ -8,6 +8,7 @@ class PlatformEnum(enum.Enum):
     CUDA = enum.auto()
     ROCM = enum.auto()
     TPU = enum.auto()
+    XPU = enum.auto()
     CPU = enum.auto()
     UNSPECIFIED = enum.auto()
 
@@ -41,6 +42,9 @@ class Platform:
     def is_tpu(self) -> bool:
         return self._enum == PlatformEnum.TPU
 
+    def is_xpu(self) -> bool:
+        return self._enum == PlatformEnum.XPU
+
     def is_cpu(self) -> bool:
         return self._enum == PlatformEnum.CPU
 

vllm/platforms/xpu.py

+import torch
+
+from .interface import DeviceCapability, Platform, PlatformEnum
+
+
+class XPUPlatform(Platform):
+    _enum = PlatformEnum.XPU
+
+    @staticmethod
+    def get_device_capability(device_id: int = 0) -> DeviceCapability:
+        return DeviceCapability(major=int(
+            torch.xpu.get_device_capability(device_id)['version'].split('.')
+            [0]),
+                                minor=int(
+                                    torch.xpu.get_device_capability(device_id)
+                                    ['version'].split('.')[1]))
+
+    @staticmethod
+    def get_device_name(device_id: int = 0) -> str:
+        return torch.xpu.get_device_name(device_id)