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

[UX] Improve UX of CPU backend (#36968)

Browse Source 
Signed-off-by: jiang1.li <jiang1.li@intel.com>
Signed-off-by: Li, Jiang <bigpyj64@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
This commit is contained in:
Li, Jiang
2026-03-14 09:27:29 +08:00
committed by GitHub
parent f680dc1b39
commit 092ace9e3a
10 changed files with 174 additions and 118 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
docker/Dockerfile.cpu
View File

@@ -14,12 +14,7 @@
#
# Build arguments:
# PYTHON_VERSION=3.13|3.12 (default)|3.11|3.10
# VLLM_CPU_DISABLE_AVX512=false (default)|true
# VLLM_CPU_AVX2=false (default)|true (for cross-compilation)
# VLLM_CPU_AVX512=false (default)|true (for cross-compilation)
# VLLM_CPU_AVX512BF16=false (default)|true (for cross-compilation)
# VLLM_CPU_AVX512VNNI=false (default)|true (for cross-compilation)
# VLLM_CPU_AMXBF16=false (default)|true (for cross-compilation)
# VLLM_CPU_X86=false (default)|true (for cross-compilation)
# VLLM_CPU_ARM_BF16=false (default)|true (for cross-compilation)
#
@@ -36,7 +31,7 @@ RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
--mount=type=cache,target=/var/lib/apt,sharing=locked \
apt-get update -y \
&& apt-get install -y --no-install-recommends sudo ccache git curl wget ca-certificates \
gcc-12 g++-12 libtcmalloc-minimal4 libnuma-dev ffmpeg libsm6 libxext6 libgl1 jq lsof \
gcc-12 g++-12 libtcmalloc-minimal4 libnuma-dev ffmpeg libsm6 libxext6 libgl1 jq lsof xz-utils \
&& update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-12 10 --slave /usr/bin/g++ g++ /usr/bin/g++-12 \
&& curl -LsSf https://astral.sh/uv/install.sh | sh
@@ -91,24 +86,9 @@ ARG max_jobs=32
ENV MAX_JOBS=${max_jobs}
ARG GIT_REPO_CHECK=0
# Support for building with non-AVX512 vLLM: docker build --build-arg VLLM_CPU_DISABLE_AVX512="true" ...
ARG VLLM_CPU_DISABLE_AVX512=0
ENV VLLM_CPU_DISABLE_AVX512=${VLLM_CPU_DISABLE_AVX512}
# Support for cross-compilation with AVX2 ISA: docker build --build-arg VLLM_CPU_AVX2="1" ...
ARG VLLM_CPU_AVX2=0
ENV VLLM_CPU_AVX2=${VLLM_CPU_AVX2}
# Support for cross-compilation with AVX512 ISA: docker build --build-arg VLLM_CPU_AVX512="1" ...
ARG VLLM_CPU_AVX512=0
ENV VLLM_CPU_AVX512=${VLLM_CPU_AVX512}
# Support for building with AVX512BF16 ISA: docker build --build-arg VLLM_CPU_AVX512BF16="true" ...
ARG VLLM_CPU_AVX512BF16=0
ENV VLLM_CPU_AVX512BF16=${VLLM_CPU_AVX512BF16}
# Support for building with AVX512VNNI ISA: docker build --build-arg VLLM_CPU_AVX512VNNI="true" ...
ARG VLLM_CPU_AVX512VNNI=0
ENV VLLM_CPU_AVX512VNNI=${VLLM_CPU_AVX512VNNI}
# Support for building with AMXBF16 ISA: docker build --build-arg VLLM_CPU_AMXBF16="true" ...
ARG VLLM_CPU_AMXBF16=1
ENV VLLM_CPU_AMXBF16=${VLLM_CPU_AMXBF16}
# Support for cross-compilation with x86 ISA including AVX2 and AVX512: docker build --build-arg VLLM_CPU_X86="true" ...
ARG VLLM_CPU_X86=0
ENV VLLM_CPU_X86=${VLLM_CPU_X86}
# Support for cross-compilation with ARM BF16 ISA: docker build --build-arg VLLM_CPU_ARM_BF16="true" ...
ARG VLLM_CPU_ARM_BF16=0
ENV VLLM_CPU_ARM_BF16=${VLLM_CPU_ARM_BF16}
@@ -116,7 +96,7 @@ ENV VLLM_CPU_ARM_BF16=${VLLM_CPU_ARM_BF16}
WORKDIR /vllm-workspace
# Validate build arguments - prevent mixing incompatible ISA flags
RUN if [ "$TARGETARCH" = "arm64" ] && { [ "$VLLM_CPU_AVX2" != "0" ] || [ "$VLLM_CPU_AVX512" != "0" ] || [ "$VLLM_CPU_AVX512BF16" != "0" ] || [ "$VLLM_CPU_AVX512VNNI" != "0" ]; }; then \
RUN if [ "$TARGETARCH" = "arm64" ] && [ "$VLLM_CPU_X86" != "0" ]; then \
echo "ERROR: Cannot use x86-specific ISA flags (AVX2, AVX512, etc.) when building for ARM64 (--platform=linux/arm64)"; \
exit 1; \
fi && \
@@ -174,7 +154,7 @@ WORKDIR /vllm-workspace
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
--mount=type=cache,target=/var/lib/apt,sharing=locked \
apt-get install -y --no-install-recommends vim numactl xz-utils make clangd-14
apt-get install -y --no-install-recommends vim numactl make clangd-14
RUN ln -s /usr/bin/clangd-14 /usr/bin/clangd
@@ -232,22 +212,12 @@ LABEL org.opencontainers.image.source="https://github.com/vllm-project/vllm"
# Build configuration labels
ARG TARGETARCH
ARG VLLM_CPU_DISABLE_AVX512
ARG VLLM_CPU_AVX2
ARG VLLM_CPU_AVX512
ARG VLLM_CPU_AVX512BF16
ARG VLLM_CPU_AVX512VNNI
ARG VLLM_CPU_AMXBF16
ARG VLLM_CPU_X86
ARG VLLM_CPU_ARM_BF16
ARG PYTHON_VERSION
LABEL ai.vllm.build.target-arch="${TARGETARCH}"
LABEL ai.vllm.build.cpu-disable-avx512="${VLLM_CPU_DISABLE_AVX512:-false}"
LABEL ai.vllm.build.cpu-avx2="${VLLM_CPU_AVX2:-false}"
LABEL ai.vllm.build.cpu-avx512="${VLLM_CPU_AVX512:-false}"
LABEL ai.vllm.build.cpu-avx512bf16="${VLLM_CPU_AVX512BF16:-false}"
LABEL ai.vllm.build.cpu-avx512vnni="${VLLM_CPU_AVX512VNNI:-false}"
LABEL ai.vllm.build.cpu-amxbf16="${VLLM_CPU_AMXBF16:-false}"
LABEL ai.vllm.build.cpu-x86="${VLLM_CPU_X86:-false}"
LABEL ai.vllm.build.cpu-arm-bf16="${VLLM_CPU_ARM_BF16:-false}"
LABEL ai.vllm.build.python-version="${PYTHON_VERSION:-3.12}"