MSCCL++ Integration (#22734)

Co-authored-by: Caio Rocha <caiorocha@microsof.com>
Co-authored-by: empyreus <rjsouza1995@gmail.com>
This commit is contained in:
Caio Rocha
2026-06-08 21:13:13 -07:00
committed by GitHub
co-authored by Caio Rocha empyreus
parent 9c53031d2b
commit c2eae96c56
17 changed files with 397 additions and 1775 deletions
+20
View File
@@ -21,6 +21,7 @@ ARG GITHUB_ARTIFACTORY=github.com
ARG INSTALL_FLASHINFER_JIT_CACHE=0
ARG FLASHINFER_VERSION=0.6.12
ARG MOONCAKE_VERSION=0.3.11.post1
ARG MSCCLPP_VERSION=sglang-v0.9.1
#if need other arg please add in MOONCAKE_COMPILE_ARG
ARG MOONCAKE_COMPILE_ARG="-DUSE_HTTP=ON -DUSE_MNNVL=ON -DUSE_CUDA=ON -DWITH_EP=ON"
@@ -443,6 +444,7 @@ ARG USE_LATEST_SGLANG
ARG GITHUB_ARTIFACTORY
ARG MOONCAKE_VERSION
ARG MOONCAKE_COMPILE_ARG
ARG MSCCLPP_VERSION
WORKDIR /sgl-workspace
@@ -523,6 +525,24 @@ RUN --mount=type=cache,target=/root/.cache/pip \
python3 -m pip install mooncake-transfer-engine==${MOONCAKE_VERSION}; \
fi
# 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]"; \
;; \
*) \
echo "Unsupported CUDA version for MSCCL++: ${CUDA_VERSION}" && exit 1; \
;; \
esac \
&& rm -rf /tmp/mscclpp
# Install essential Python packages (use constraints to prevent conflicts)
RUN --mount=type=cache,target=/root/.cache/pip \
python3 -m pip install -c /sgl-workspace/constraints.txt \