### What this PR does / why we need it? Triton-ascend occasionally encounters compilation errors, which is a known issue in triton-ascend 3.2.0. However, we want to use the official version rather than the development version, so we only changed the triton-ascend version in the Dockerfile and added a FAQ to explain this issue. --------- Signed-off-by: MengqingCao <cmq0113@163.com>
88 lines
4.3 KiB
Docker
88 lines
4.3 KiB
Docker
#
|
|
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
# This file is a part of the vllm-ascend project.
|
|
#
|
|
|
|
FROM quay.io/ascend/cann:8.5.1-a3-openeuler24.03-py3.11
|
|
|
|
ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
|
|
ARG MOONCAKE_TAG="v0.3.8.post1"
|
|
ARG SOC_VERSION="ascend910_9391"
|
|
|
|
ENV SOC_VERSION=$SOC_VERSION \
|
|
TASK_QUEUE_ENABLE=1 \
|
|
OMP_NUM_THREADS=1
|
|
|
|
RUN pip config set global.index-url ${PIP_INDEX_URL}
|
|
|
|
WORKDIR /workspace
|
|
|
|
COPY . /vllm-workspace/vllm-ascend/
|
|
|
|
SHELL ["/bin/bash", "-c"]
|
|
|
|
RUN yum update -y && \
|
|
yum install -y git vim wget net-tools gcc gcc-c++ make cmake numactl-devel jemalloc && \
|
|
git clone --depth 1 --branch ${MOONCAKE_TAG} https://github.com/kvcache-ai/Mooncake /vllm-workspace/Mooncake && \
|
|
cp /vllm-workspace/vllm-ascend/tools/mooncake_installer.sh /vllm-workspace/Mooncake/ && \
|
|
ARCH=$(uname -m) && \
|
|
source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
|
|
export LD_LIBRARY_PATH=/usr/local/Ascend/ascend-toolkit/latest/${ARCH}-linux/devlib:/usr/local/Ascend/ascend-toolkit/latest/${ARCH}-linux/lib64:$LD_LIBRARY_PATH && \
|
|
export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/${ARCH}-openEuler-linux && \
|
|
cd /vllm-workspace/Mooncake && \
|
|
bash mooncake_installer.sh -y && \
|
|
mkdir -p build && cd build && cmake .. -DUSE_ASCEND_DIRECT=ON && \
|
|
make -j$(nproc) && make install && \
|
|
rm -fr /vllm-workspace/Mooncake/build && \
|
|
rm -rf /var/cache/yum/*
|
|
|
|
# Install vLLM
|
|
ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
|
|
ARG VLLM_TAG=v0.18.0
|
|
RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /vllm-workspace/vllm
|
|
# In x86, triton will be installed by vllm. But in Ascend, triton doesn't work correctly. we need to uninstall it.
|
|
RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -e /vllm-workspace/vllm/[audio] --extra-index https://download.pytorch.org/whl/cpu/ && \
|
|
python3 -m pip uninstall -y triton && \
|
|
python3 -m pip cache purge
|
|
|
|
# Install vllm-ascend
|
|
# Installing vllm-ascend on x86 can pull upstream triton back in alongside triton-ascend. Remove it immediately after this step.
|
|
RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
|
|
source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
|
|
source /usr/local/Ascend/nnal/atb/set_env.sh && \
|
|
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
|
|
export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
|
|
python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
|
|
if [ "$(uname -i)" = "x86_64" ]; then python3 -m pip uninstall -y triton; fi && \
|
|
PYTHON_TAG=$(python3 -c "import sys; print(f'cp{sys.version_info.major}{sys.version_info.minor}')") && \
|
|
ARCH=$(python3 -c "import platform; machine = platform.machine().lower(); arch_map = {'x86_64': 'x86_64', 'amd64': 'x86_64', 'aarch64': 'aarch64', 'arm64': 'aarch64'}; print(arch_map.get(machine, machine))") && \
|
|
TRITON_ASCEND_WHEEL="triton_ascend-3.2.0.dev20260322-${PYTHON_TAG}-${PYTHON_TAG}-manylinux_2_27_${ARCH}.manylinux_2_28_${ARCH}.whl" && \
|
|
python3 -m pip install "https://vllm-ascend.obs.cn-north-4.myhuaweicloud.com/vllm-ascend/${TRITON_ASCEND_WHEEL}" && \
|
|
python3 -m pip cache purge
|
|
|
|
# Install clang (for triton-ascend)
|
|
RUN yum update -y && \
|
|
yum install -y clang && \
|
|
rm -rf /var/cache/yum/*
|
|
|
|
# Install modelscope (for fast download) and ray (for multinode)
|
|
RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
|
|
python3 -m pip cache purge
|
|
|
|
RUN echo "export LD_PRELOAD=/usr/lib64/libjemalloc.so.2:$LD_PRELOAD" >> ~/.bashrc
|
|
RUN echo "export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib" >> ~/.bashrc
|
|
|
|
CMD ["/bin/bash"]
|