hailin
/
vllm
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

.

This commit is contained in:
hailin 2025-08-05 19:57:11 +08:00
parent f2048d004f
commit e9abf3cb03
1 changed files with 43 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

65
Dockerfile
View File

@ -5,7 +5,6 @@ ARG PYTHON_VERSION=3.12
ARG BUILD_BASE_IMAGE=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu20.04
ARG FINAL_BASE_IMAGE=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu22.04
# ARG DEADSNAKES_MIRROR_URL
# ARG DEADSNAKES_GPGKEY_URL
@ -108,15 +107,29 @@ RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install --system -r requirements/cuda.txt \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
ARG torch_cuda_arch_list='7.0 7.5 8.0 8.9 9.0 10.0 12.0'
# ARG torch_cuda_arch_list='7.0 7.5 8.0 8.9 9.0 10.0 12.0'
ARG torch_cuda_arch_list='8.6 8.9'
ENV TORCH_CUDA_ARCH_LIST=${torch_cuda_arch_list}
# Override the arch list for flash-attn to reduce the binary size
ARG vllm_fa_cmake_gpu_arches='80-real;90-real'
# ARG vllm_fa_cmake_gpu_arches='80-real;90-real'
ARG vllm_fa_cmake_gpu_arches='86-real;89-real'
ENV VLLM_FA_CMAKE_GPU_ARCHES=${vllm_fa_cmake_gpu_arches}
#################### BASE BUILD IMAGE ####################
#################### WHEEL BUILD IMAGE ####################
FROM base AS build
RUN curl -L -o sccache.tar.gz https://github.com/mozilla/sccache/releases/download/v0.8.1/sccache-v0.8.1-x86_64-unknown-linux-musl.tar.gz && \
tar -xzf sccache.tar.gz && mv sccache-*/sccache /usr/local/bin/sccache && rm -rf sccache*
ENV SCCACHE_DIR=/home/jzy/.cache/sccache
ENV SCCACHE_IDLE_TIMEOUT=0
ENV RUSTC_WRAPPER=sccache
ENV CMAKE_C_COMPILER_LAUNCHER=sccache
ENV CMAKE_CXX_COMPILER_LAUNCHER=sccache
ENV CMAKE_CUDA_COMPILER_LAUNCHER=sccache
ARG TARGETPLATFORM
ARG PIP_INDEX_URL UV_INDEX_URL
@ -165,24 +178,31 @@ RUN if [ "${VLLM_USE_PRECOMPILED}" = "1" ]; then \
fi
# if USE_SCCACHE is set, use sccache to speed up compilation
RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=bind,source=.git,target=.git \
if [ "$USE_SCCACHE" = "1" ]; then \
echo "Installing sccache..." \
&& curl -L -o sccache.tar.gz ${SCCACHE_DOWNLOAD_URL} \
&& tar -xzf sccache.tar.gz \
&& sudo mv sccache-v0.8.1-x86_64-unknown-linux-musl/sccache /usr/bin/sccache \
&& rm -rf sccache.tar.gz sccache-v0.8.1-x86_64-unknown-linux-musl \
&& if [ ! -z ${SCCACHE_ENDPOINT} ] ; then export SCCACHE_ENDPOINT=${SCCACHE_ENDPOINT} ; fi \
&& export SCCACHE_BUCKET=${SCCACHE_BUCKET_NAME} \
&& export SCCACHE_REGION=${SCCACHE_REGION_NAME} \
&& export SCCACHE_S3_NO_CREDENTIALS=${SCCACHE_S3_NO_CREDENTIALS} \
&& export SCCACHE_IDLE_TIMEOUT=0 \
&& export CMAKE_BUILD_TYPE=Release \
&& sccache --show-stats \
&& python3 setup.py bdist_wheel --dist-dir=dist --py-limited-api=cp38 \
&& sccache --show-stats; \
fi
# RUN --mount=type=cache,target=/root/.cache/uv \
# --mount=type=bind,source=.git,target=.git \
# if [ "$USE_SCCACHE" = "1" ]; then \
# echo "Installing sccache..." \
# && curl -L -o sccache.tar.gz ${SCCACHE_DOWNLOAD_URL} \
# && tar -xzf sccache.tar.gz \
# && sudo mv sccache-v0.8.1-x86_64-unknown-linux-musl/sccache /usr/bin/sccache \
# && rm -rf sccache.tar.gz sccache-v0.8.1-x86_64-unknown-linux-musl \
# && if [ ! -z ${SCCACHE_ENDPOINT} ] ; then export SCCACHE_ENDPOINT=${SCCACHE_ENDPOINT} ; fi \
# && export SCCACHE_BUCKET=${SCCACHE_BUCKET_NAME} \
# && export SCCACHE_REGION=${SCCACHE_REGION_NAME} \
# && export SCCACHE_S3_NO_CREDENTIALS=${SCCACHE_S3_NO_CREDENTIALS} \
# && export SCCACHE_IDLE_TIMEOUT=0 \
# && export CMAKE_BUILD_TYPE=Release \
# && sccache --show-stats \
# && python3 setup.py bdist_wheel --dist-dir=dist --py-limited-api=cp38 \
# && sccache --show-stats; \
# fi
RUN --mount=type=cache,target=/root/.cache/sccache \
--mount=type=cache,target=/root/.cache/uv \
sccache --start-server && \
sccache --zero-stats && \
python3 setup.py bdist_wheel --dist-dir=dist --py-limited-api=cp38 && \
sccache --show-stats
ENV CCACHE_DIR=/root/.cache/ccache
RUN --mount=type=cache,target=/root/.cache/ccache \
@ -325,7 +345,8 @@ RUN --mount=type=cache,target=/root/.cache/uv bash - <<'BASH'
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a"
else
# CUDA 12.8+ supports 10.0a and 12.0
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a 10.0a 12.0"
# FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a 10.0a 12.0"
FI_TORCH_CUDA_ARCH_LIST="8.6 8.9"
fi
echo "🏗️ Building FlashInfer for arches: ${FI_TORCH_CUDA_ARCH_LIST}"
# Needed to build AOT kernels