[ROCm][CI/Build] Sync ROCm dockerfiles with the ROCm fork (vllm-project#24279)

Signed-off-by: Gregory Shtrasberg <[email protected]>

Author: gshtras

docker/Dockerfile.rocm

@@ -47,6 +47,7 @@ COPY --from=build_vllm ${COMMON_WORKDIR}/vllm/requirements /requirements
 COPY --from=build_vllm ${COMMON_WORKDIR}/vllm/benchmarks /benchmarks
 COPY --from=build_vllm ${COMMON_WORKDIR}/vllm/tests /tests
 COPY --from=build_vllm ${COMMON_WORKDIR}/vllm/examples /examples
+COPY --from=build_vllm ${COMMON_WORKDIR}/vllm/docker/Dockerfile.rocm /docker/
 COPY --from=build_vllm ${COMMON_WORKDIR}/vllm/.buildkite /.buildkite
 
 # -----------------------
@@ -71,7 +72,7 @@ COPY --from=build_vllm ${COMMON_WORKDIR}/vllm /vllm-workspace
 RUN cd /vllm-workspace \
     && rm -rf vllm \
     && python3 -m pip install -e tests/vllm_test_utils \
-    && python3 -m pip install git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d#egg=lm-eval[api] \
+    && python3 -m pip install lm-eval[api]==0.4.4 \
     && python3 -m pip install pytest-shard
 
 # -----------------------
@@ -100,6 +101,7 @@ ARG COMMON_WORKDIR
 # Copy over the benchmark scripts as well
 COPY --from=export_vllm /benchmarks ${COMMON_WORKDIR}/vllm/benchmarks
 COPY --from=export_vllm /examples ${COMMON_WORKDIR}/vllm/examples
+COPY --from=export_vllm /docker ${COMMON_WORKDIR}/vllm/docker
 
 ENV RAY_EXPERIMENTAL_NOSET_ROCR_VISIBLE_DEVICES=1
 ENV TOKENIZERS_PARALLELISM=false

docker/Dockerfile.rocm_base

@@ -1,18 +1,16 @@
-ARG BASE_IMAGE=rocm/dev-ubuntu-22.04:6.3.1-complete
-ARG HIPBLASLT_BRANCH="db8e93b4"
-ARG HIPBLAS_COMMON_BRANCH="7c1566b"
+ARG BASE_IMAGE=rocm/dev-ubuntu-22.04:6.4.1-complete
+ARG HIPBLASLT_BRANCH="aa0bda7b"
+ARG HIPBLAS_COMMON_BRANCH="9b80ba8e"
 ARG LEGACY_HIPBLASLT_OPTION=
-ARG RCCL_BRANCH="648a58d"
-ARG RCCL_REPO="https://github.com/ROCm/rccl"
 ARG TRITON_BRANCH="e5be006"
 ARG TRITON_REPO="https://github.com/triton-lang/triton.git"
-ARG PYTORCH_BRANCH="295f2ed4"
+ARG PYTORCH_BRANCH="f717b2af"
 ARG PYTORCH_VISION_BRANCH="v0.21.0"
-ARG PYTORCH_REPO="https://github.com/pytorch/pytorch.git"
+ARG PYTORCH_REPO="https://github.com/ROCm/pytorch.git"
 ARG PYTORCH_VISION_REPO="https://github.com/pytorch/vision.git"
 ARG FA_BRANCH="1a7f4dfa"
 ARG FA_REPO="https://github.com/Dao-AILab/flash-attention.git"
-ARG AITER_BRANCH="916bf3c"
+ARG AITER_BRANCH="4822e675"
 ARG AITER_REPO="https://github.com/ROCm/aiter.git"
 
 FROM ${BASE_IMAGE} AS base
@@ -45,14 +43,15 @@ RUN apt-get update -y \
     && curl -sS https://bootstrap.pypa.io/get-pip.py | python${PYTHON_VERSION} \
     && python3 --version && python3 -m pip --version
 
-RUN pip install -U packaging 'cmake<4' ninja wheel setuptools pybind11 Cython
+RUN pip install -U packaging 'cmake<4' ninja wheel 'setuptools<80' pybind11 Cython
 
 FROM base AS build_hipblaslt
 ARG HIPBLASLT_BRANCH
 ARG HIPBLAS_COMMON_BRANCH
 # Set to "--legacy_hipblas_direct" for ROCm<=6.2
 ARG LEGACY_HIPBLASLT_OPTION
 RUN git clone https://github.com/ROCm/hipBLAS-common.git
+RUN apt-get remove -y hipblaslt && apt-get autoremove -y && apt-get autoclean -y
 RUN cd hipBLAS-common \
     && git checkout ${HIPBLAS_COMMON_BRANCH} \
     && mkdir build \
@@ -69,24 +68,17 @@ RUN cd hipBLASLt \
     && make package
 RUN mkdir -p /app/install && cp /app/hipBLASLt/build/release/*.deb /app/hipBLAS-common/build/*.deb /app/install
 
-FROM base AS build_rccl
-ARG RCCL_BRANCH
-ARG RCCL_REPO
-RUN git clone ${RCCL_REPO}
-RUN cd rccl \
-    && git checkout ${RCCL_BRANCH} \
-    && ./install.sh -p --amdgpu_targets ${PYTORCH_ROCM_ARCH}
-RUN mkdir -p /app/install && cp /app/rccl/build/release/*.deb /app/install
-
 FROM base AS build_triton
 ARG TRITON_BRANCH
 ARG TRITON_REPO
 RUN git clone ${TRITON_REPO}
 RUN cd triton \
     && git checkout ${TRITON_BRANCH} \
-    && cd python \
-    && python3 setup.py bdist_wheel --dist-dir=dist
-RUN mkdir -p /app/install && cp /app/triton/python/dist/*.whl /app/install
+    && if [ ! -f setup.py ]; then cd python; fi \
+    && python3 setup.py bdist_wheel --dist-dir=dist \
+    && mkdir -p /app/install && cp dist/*.whl /app/install
+RUN if [ -d triton/python/triton_kernels ]; then pip install build && cd triton/python/triton_kernels \
+    && python3 -m build --wheel && cp dist/*.whl /app/install; fi
 
 FROM base AS build_amdsmi
 RUN cd /opt/rocm/share/amd_smi \
@@ -132,15 +124,25 @@ RUN cd aiter \
 RUN pip install pyyaml && cd aiter && PREBUILD_KERNELS=1 GPU_ARCHS=gfx942 python3 setup.py bdist_wheel --dist-dir=dist && ls /app/aiter/dist/*.whl
 RUN mkdir -p /app/install && cp /app/aiter/dist/*.whl /app/install
 
+FROM base AS debs
+RUN mkdir /app/debs
+RUN --mount=type=bind,from=build_hipblaslt,src=/app/install/,target=/install \
+    cp /install/*.deb /app/debs
+RUN --mount=type=bind,from=build_triton,src=/app/install/,target=/install \
+    cp /install/*.whl /app/debs
+RUN --mount=type=bind,from=build_amdsmi,src=/app/install/,target=/install \
+    cp /install/*.whl /app/debs
+RUN --mount=type=bind,from=build_pytorch,src=/app/install/,target=/install \
+    cp /install/*.whl /app/debs
+RUN --mount=type=bind,from=build_aiter,src=/app/install/,target=/install \
+    cp /install/*.whl /app/debs
+
 FROM base AS final
 RUN --mount=type=bind,from=build_hipblaslt,src=/app/install/,target=/install \
     dpkg -i /install/*deb \
-    && sed -i 's/, hipblaslt-dev \(.*\), hipcub-dev/, hipcub-dev/g' /var/lib/dpkg/status \
-    && sed -i 's/, hipblaslt \(.*\), hipfft/, hipfft/g' /var/lib/dpkg/status
-RUN --mount=type=bind,from=build_rccl,src=/app/install/,target=/install \
-    dpkg -i /install/*deb \
-    && sed -i 's/, rccl-dev \(.*\), rocalution/, rocalution/g' /var/lib/dpkg/status \
-    && sed -i 's/, rccl \(.*\), rocalution/, rocalution/g' /var/lib/dpkg/status
+    && perl -p -i -e 's/, hipblas-common-dev \([^)]*?\), /, /g' /var/lib/dpkg/status \
+    && perl -p -i -e 's/, hipblaslt-dev \([^)]*?\), /, /g' /var/lib/dpkg/status \
+    && perl -p -i -e 's/, hipblaslt \([^)]*?\), /, /g' /var/lib/dpkg/status
 RUN --mount=type=bind,from=build_triton,src=/app/install/,target=/install \
     pip install /install/*.whl
 RUN --mount=type=bind,from=build_amdsmi,src=/app/install/,target=/install \
@@ -154,8 +156,6 @@ ARG BASE_IMAGE
 ARG HIPBLAS_COMMON_BRANCH
 ARG HIPBLASLT_BRANCH
 ARG LEGACY_HIPBLASLT_OPTION
-ARG RCCL_BRANCH
-ARG RCCL_REPO
 ARG TRITON_BRANCH
 ARG TRITON_REPO
 ARG PYTORCH_BRANCH
@@ -170,8 +170,6 @@ RUN echo "BASE_IMAGE: ${BASE_IMAGE}" > /app/versions.txt \
     && echo "HIPBLAS_COMMON_BRANCH: ${HIPBLAS_COMMON_BRANCH}" >> /app/versions.txt \
     && echo "HIPBLASLT_BRANCH: ${HIPBLASLT_BRANCH}" >> /app/versions.txt \
     && echo "LEGACY_HIPBLASLT_OPTION: ${LEGACY_HIPBLASLT_OPTION}" >> /app/versions.txt \
-    && echo "RCCL_BRANCH: ${RCCL_BRANCH}" >> /app/versions.txt \
-    && echo "RCCL_REPO: ${RCCL_REPO}" >> /app/versions.txt \
     && echo "TRITON_BRANCH: ${TRITON_BRANCH}" >> /app/versions.txt \
     && echo "TRITON_REPO: ${TRITON_REPO}" >> /app/versions.txt \
     && echo "PYTORCH_BRANCH: ${PYTORCH_BRANCH}" >> /app/versions.txt \
@@ -180,4 +178,4 @@ RUN echo "BASE_IMAGE: ${BASE_IMAGE}" > /app/versions.txt \
     && echo "PYTORCH_VISION_REPO: ${PYTORCH_VISION_REPO}" >> /app/versions.txt \
     && echo "FA_BRANCH: ${FA_BRANCH}" >> /app/versions.txt \
     && echo "AITER_BRANCH: ${AITER_BRANCH}" >> /app/versions.txt \
-    && echo "AITER_REPO: ${AITER_REPO}" >> /app/versions.txt
+    && echo "AITER_REPO: ${AITER_REPO}" >> /app/versions.txt