[Deps] Retire the CUDA 12 lane (#38404)
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai>
This commit is contained in:
Mohammad Miadh Angkad
co-authored by
Mohammad Angkad
Mohammad Angkad
parent
fae8cd84cb
commit
52c191da52
+13
-68
@@ -493,19 +493,11 @@ RUN curl --proto '=https' --tlsv1.2 --retry 3 --retry-delay 2 -sSf https://sh.ru
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
python3 -m pip install --upgrade pip setuptools wheel html5lib six \
|
||||
&& case "$CUDA_VERSION" in \
|
||||
12.6.3) CUINDEX=126 ;; \
|
||||
12.9.2) CUINDEX=129 ;; \
|
||||
13.0.3) CUINDEX=130 ;; \
|
||||
13.4.0) CUINDEX=134 ;; \
|
||||
*) echo "Unsupported CUDA version: $CUDA_VERSION" && exit 1 ;; \
|
||||
esac \
|
||||
&& if [ "$CUDA_VERSION" = "12.6.3" ]; then \
|
||||
python3 -m pip install https://${GITHUB_ARTIFACTORY}/sgl-project/whl/releases/download/v${SGL_KERNEL_VERSION}/sglang_kernel-${SGL_KERNEL_VERSION}+cu124-cp310-abi3-manylinux2014_$(uname -m).whl --force-reinstall --no-deps \
|
||||
; \
|
||||
elif [ "$CUDA_VERSION" = "12.9.2" ]; then \
|
||||
python3 -m pip install https://${GITHUB_ARTIFACTORY}/sgl-project/whl/releases/download/v${SGL_KERNEL_VERSION}/sglang_kernel-${SGL_KERNEL_VERSION}+cu129-cp310-abi3-manylinux2014_$(uname -m).whl --force-reinstall --no-deps \
|
||||
; \
|
||||
elif [ "$CUDA_VERSION" = "13.0.3" ]; then \
|
||||
&& if [ "$CUDA_VERSION" = "13.0.3" ]; then \
|
||||
# --no-deps prevents pip from pulling torch from default PyPI
|
||||
python3 -m pip install sglang-kernel==${SGL_KERNEL_VERSION} --force-reinstall --no-deps \
|
||||
; \
|
||||
@@ -527,15 +519,11 @@ COPY rust/sglang-grpc /tmp/sglang_deps/rust/sglang-grpc
|
||||
COPY rust/sglang-mm /tmp/sglang_deps/rust/sglang-mm
|
||||
COPY proto /tmp/sglang_deps/proto
|
||||
|
||||
# Install sglang dependencies (torch, transformers, etc.). CUDA 12 DeepEP
|
||||
# wheels live only on the SGLang index, so preinstall the local-version wheel;
|
||||
# it satisfies the public-version pyproject pin during the full dependency solve.
|
||||
# Install sglang dependencies (torch, transformers, etc.).
|
||||
# Generate constraints.txt to prevent reinstalling these deps in later stages.
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
--mount=type=cache,target=/root/.cargo/registry \
|
||||
case "$CUDA_VERSION" in \
|
||||
12.6.3) CUINDEX=126 ;; \
|
||||
12.9.2) CUINDEX=129 ;; \
|
||||
13.0.3) CUINDEX=130 ;; \
|
||||
13.4.0) CUINDEX=134 ;; \
|
||||
*) echo "Unsupported CUDA version: $CUDA_VERSION" && exit 1 ;; \
|
||||
@@ -546,19 +534,6 @@ RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
&& echo '__version__ = "0.0.0"' > sglang/version.py \
|
||||
&& touch README.md \
|
||||
&& touch LICENSE \
|
||||
&& SGL_DEEP_EP_VERSION="$(sed -n 's/^[[:space:]]*"sgl-deep-ep==\([^"]*\)",/\1/p' pyproject.toml)" \
|
||||
&& test -n "${SGL_DEEP_EP_VERSION}" \
|
||||
&& if [ "${CUDA_VERSION%%.*}" = "12" ]; then \
|
||||
python3 -m pip install \
|
||||
"sgl-deep-ep==${SGL_DEEP_EP_VERSION}+cu129" \
|
||||
--index-url "https://docs.sglang.ai/whl/cu129/" \
|
||||
--no-deps; \
|
||||
fi \
|
||||
&& if [ "${CUDA_VERSION%%.*}" = "12" ]; then \
|
||||
sed -i 's/cuda-python>=13\.0/cuda-python>=12,<13/' pyproject.toml && \
|
||||
sed -i 's/flashinfer_python\[cu13\]/flashinfer_python[cu12]/' pyproject.toml && \
|
||||
sed -i 's/nvidia-cutlass-dsl\[cu13\]/nvidia-cutlass-dsl/' pyproject.toml; \
|
||||
fi \
|
||||
&& if [ "$CUDA_VERSION" = "13.4.0" ]; then \
|
||||
sed -i "s|^ \"flashinfer_python\[cu13\]==.*| \"flashinfer_python[cu13]==${FLASHINFER_VERSION}\",|" pyproject.toml; \
|
||||
fi \
|
||||
@@ -579,17 +554,8 @@ RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
TORCH_INDEX="https://download.pytorch.org/whl/cu${CUINDEX}"; \
|
||||
fi \
|
||||
&& python3 -m pip install --extra-index-url "${TORCH_INDEX}" ".[${BUILD_TYPE}]" \
|
||||
&& if [ "${CUDA_VERSION%%.*}" = "13" ]; then \
|
||||
python3 -m pip install --force-reinstall --no-deps \
|
||||
"nvidia-nccl-cu13==${NCCL_VERSION}"; \
|
||||
fi \
|
||||
&& if [ "${CUDA_VERSION%%.*}" = "12" ]; then \
|
||||
pip list --format=freeze | awk -F'==' '/-cu13(==|$)/ {print $1}' \
|
||||
| xargs -r python3 -m pip uninstall -y && \
|
||||
python3 -m pip install --index-url https://download.pytorch.org/whl/cu${CUINDEX} \
|
||||
torch==2.13.0 torchvision==0.28.0 torchaudio==2.11.0 --force-reinstall; \
|
||||
python3 -m pip install https://github.com/sgl-project/whl/releases/download/v${SGL_DEEP_GEMM_VERSION}/sgl_deep_gemm-${SGL_DEEP_GEMM_VERSION}+cu129-py3-none-manylinux2014_$(uname -m).whl --force-reinstall; \
|
||||
fi \
|
||||
&& python3 -m pip install --force-reinstall --no-deps \
|
||||
"nvidia-nccl-cu13==${NCCL_VERSION}" \
|
||||
&& if [ "$CUDA_VERSION" = "13.4.0" ]; then \
|
||||
python3 -m pip install --force-reinstall --no-deps /tmp/cu134_wheels/*.whl; \
|
||||
fi \
|
||||
@@ -643,8 +609,6 @@ ARG FLASHINFER_VERSION
|
||||
# Stage jit-cache/cubin artifacts into /flashinfer_jit_output for clean COPY later
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
case "$CUDA_VERSION" in \
|
||||
12.6.3) CUINDEX=126 ;; \
|
||||
12.9.2) CUINDEX=129 ;; \
|
||||
13.0.3) CUINDEX=130 ;; \
|
||||
13.4.0) CUINDEX=134 ;; \
|
||||
*) echo "Unsupported CUDA version: $CUDA_VERSION" && exit 1 ;; \
|
||||
@@ -830,21 +794,12 @@ RUN --mount=type=cache,target=/var/cache/apt,id=framework-apt \
|
||||
|
||||
# Install Mooncake
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
CUDA_MAJOR="${CUDA_VERSION%%.*}" && \
|
||||
if [ "$CUDA_MAJOR" -ge 13 ]; then \
|
||||
python3 -m pip install mooncake-transfer-engine-cuda13==${MOONCAKE_VERSION}; \
|
||||
else \
|
||||
python3 -m pip install mooncake-transfer-engine==${MOONCAKE_VERSION}; \
|
||||
fi
|
||||
python3 -m pip install mooncake-transfer-engine-cuda13==${MOONCAKE_VERSION}
|
||||
|
||||
# Install MSCCL++ Python dependencies and package (builds extension via CMake through pip)
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
git clone --depth=1 --branch ${MSCCLPP_VERSION} https://${GITHUB_ARTIFACTORY}/microsoft/mscclpp.git /tmp/mscclpp \
|
||||
&& case "${CUDA_VERSION}" in \
|
||||
12.*) \
|
||||
CMAKE_ARGS="-DMSCCLPP_BYPASS_GPU_CHECK=ON -DMSCCLPP_USE_CUDA=ON -DMSCCLPP_GPU_ARCHS=80,90,100,100a,103,103a" \
|
||||
python3 -m pip install "/tmp/mscclpp[cuda12]"; \
|
||||
;; \
|
||||
13.*) \
|
||||
CMAKE_ARGS="-DMSCCLPP_BYPASS_GPU_CHECK=ON -DMSCCLPP_USE_CUDA=ON -DMSCCLPP_GPU_ARCHS=80,90,100,100a,103,103a" \
|
||||
python3 -m pip install "/tmp/mscclpp[cuda13]"; \
|
||||
@@ -882,17 +837,12 @@ RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
termplotlib \
|
||||
"runai-model-streamer[s3,gcs,azure]>=0.15.7"
|
||||
|
||||
# Per-CUDA-major package installs. The `nixl` stub package is needed (it owns
|
||||
# the `nixl` import path) but unconditionally requires nixl-cu12, so we install
|
||||
# it with --no-deps and pair it with the matching nixl-cu12 / nixl-cu13 binary
|
||||
# to avoid shipping wrong-CUDA libs on cu13 images.
|
||||
RUN --mount=type=cache,target=/root/.cache/pip if [ "${CUDA_VERSION%%.*}" = "12" ]; then \
|
||||
python3 -m pip install nixl nixl-cu12 --no-deps ; \
|
||||
python3 -m pip install "cuda-python>=12,<13" ; \
|
||||
elif [ "${CUDA_VERSION%%.*}" = "13" ]; then \
|
||||
python3 -m pip install nixl nixl-cu13 --no-deps ; \
|
||||
python3 -m pip install "cuda-python>=13,<14" ; \
|
||||
fi
|
||||
# The `nixl` stub package is needed (it owns the `nixl` import path) but
|
||||
# unconditionally requires nixl-cu12, so install it with --no-deps and pair it
|
||||
# with the nixl-cu13 binary to avoid shipping wrong-CUDA libs.
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
python3 -m pip install nixl nixl-cu13 --no-deps \
|
||||
&& python3 -m pip install "cuda-python>=13,<14"
|
||||
|
||||
# Install the latest available AI Dynamo prerelease from NVIDIA's package index.
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
@@ -965,11 +915,6 @@ RUN if [ "$BRANCH_TYPE" = "local" ]; then \
|
||||
# Clean up __pycache__/tests/pyc in same RUN to avoid writing ~28k files to layer
|
||||
RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
cd /sgl-workspace/sglang \
|
||||
&& if [ "${CUDA_VERSION%%.*}" = "12" ]; then \
|
||||
sed -i 's/cuda-python>=13\.0/cuda-python>=12,<13/' python/pyproject.toml && \
|
||||
sed -i 's/flashinfer_python\[cu13\]/flashinfer_python[cu12]/' python/pyproject.toml && \
|
||||
sed -i 's/nvidia-cutlass-dsl\[cu13\]/nvidia-cutlass-dsl/' python/pyproject.toml; \
|
||||
fi \
|
||||
&& if [ "$CUDA_VERSION" = "13.4.0" ]; then \
|
||||
sed -i "s|^ \"flashinfer_python\[cu13\]==.*| \"flashinfer_python[cu13]==${FLASHINFER_VERSION}\",|" python/pyproject.toml; \
|
||||
fi \
|
||||
@@ -989,8 +934,8 @@ RUN --mount=type=cache,target=/root/.cache/pip \
|
||||
echo "sgl-kernel cubin download failed, retrying in 30s..." && sleep 30; \
|
||||
done; \
|
||||
# x86: if no prebuilt sgl-flash-attn3 variant matches this torch+CUDA \
|
||||
# combo (e.g. cu129 publishes no torch>=2.10 cubin), fall back to \
|
||||
# runtime JIT instead of failing the build, mirroring the aarch64 branch. \
|
||||
# combo, fall back to runtime JIT instead of failing the build, \
|
||||
# mirroring the aarch64 branch. \
|
||||
if [ "$success" != "1" ]; then \
|
||||
echo "WARNING: no matching sgl-flash-attn3 cubin variant for this torch+CUDA; kernels will be JIT-compiled at runtime"; \
|
||||
success=1; \
|
||||
|
||||
Reference in New Issue
Block a user