add env vars & misc

CMakeLists.txt

@@ -77,11 +77,12 @@ message("TORCH_NPU_PATH is ${TORCH_NPU_PATH}")
 
 if(SOC_VERSION STREQUAL "ASCEND310P3")
     file(GLOB VLLM_ASCEND_SRC
-    ${CMAKE_CURRENT_SOURCE_DIR}/csrc/*.cpp)
+    ${CMAKE_CURRENT_SOURCE_DIR}/csrc/*.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/csrc/vnpu_offload/shm_worker.cpp)
 else()
     file(GLOB VLLM_ASCEND_SRC
     ${CMAKE_CURRENT_SOURCE_DIR}/csrc/*.cpp
-    ${CMAKE_CURRENT_SOURCE_DIR}/csrc/idle_offload/shm_worker.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/csrc/vnpu_offload/shm_worker.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/csrc/batch_matmul_transpose/op_host/tiling/tiling_data.cpp)
 endif()
 
@@ -94,7 +95,7 @@ include_directories(
   ${ASCEND_HOME_PATH}/aarch64-linux/include/experiment/platform
   ${ASCEND_HOME_PATH}/x86_64-linux/include/experiment/platform
   ${CMAKE_CURRENT_SOURCE_DIR}/csrc/batch_matmul_transpose/op_host
-  ${CMAKE_CURRENT_SOURCE_DIR}/csrc/idle_offload/include
+  ${CMAKE_CURRENT_SOURCE_DIR}/csrc/vnpu_offload/include
 )
 
 set(

Dockerfile

@@ -59,14 +59,14 @@ RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
     source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
-    cd /vllm-workspace/vllm-ascend/csrc/idle_offload && \
+    cd /vllm-workspace/vllm-ascend/csrc/vnpu_offload && \
     make install && make clean && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
     python3 -m pip cache purge
 
 ENV VLLM_ASCEND_ENABLE_NZ=0 \
     VLLM_WORKER_MULTIPROC_METHOD=spawn \
-    VLLM_ASCEND_ENABLE_IDLE_OFFLOAD=1
+    VLLM_ASCEND_ENABLE_VNPU=1
 
 # Install modelscope (for fast download) and ray (for multinode)
 RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \

Dockerfile.310p

@@ -51,9 +51,15 @@ RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
     export SOC_VERSION=ASCEND310P3 && \
+    cd /vllm-workspace/vllm-ascend/csrc/vnpu_offload && \
+    make install && make clean && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
     python3 -m pip cache purge
 
+ENV VLLM_ASCEND_ENABLE_NZ=0 \
+    VLLM_WORKER_MULTIPROC_METHOD=spawn \
+    VLLM_ASCEND_ENABLE_VNPU=1
+
 # Install modelscope (for fast download) and ray (for multinode)
 RUN python3 -m pip install modelscope 'ray>=2.47.1' 'protobuf>3.20.0' && \
     python3 -m pip cache purge

Dockerfile.310p.openEuler

@@ -49,9 +49,15 @@ RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
     export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
     export SOC_VERSION=ASCEND310P3 && \
+    cd /vllm-workspace/vllm-ascend/csrc/vnpu_offload && \
+    make install && make clean && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
     python3 -m pip cache purge
 
+ENV VLLM_ASCEND_ENABLE_NZ=0 \
+    VLLM_WORKER_MULTIPROC_METHOD=spawn \
+    VLLM_ASCEND_ENABLE_VNPU=1
+
 # Install modelscope (for fast download) and ray (for multinode)
 RUN python3 -m pip install modelscope 'ray>=2.47.1' 'protobuf>3.20.0' && \
     python3 -m pip cache purge

Dockerfile.a3

@@ -20,11 +20,13 @@ FROM quay.io/ascend/cann:8.3.rc2-a3-ubuntu22.04-py3.11
 ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
 ARG COMPILE_CUSTOM_KERNELS=1
 ARG MOONCAKE_TAG=v0.3.7.post2
+ARG SOC_VERSION="ascend910_9391"
 
 COPY . /vllm-workspace/vllm-ascend/
 # Define environments
 ENV DEBIAN_FRONTEND=noninteractive
-ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS} \
+    SOC_VERSION=$SOC_VERSION
 
 RUN pip config set global.index-url ${PIP_INDEX_URL}
 
@@ -58,9 +60,15 @@ RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
     source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
+    cd /vllm-workspace/vllm-ascend/csrc/vnpu_offload && \
+    make install && make clean && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
     python3 -m pip cache purge
 
+ENV VLLM_ASCEND_ENABLE_NZ=0 \
+    VLLM_WORKER_MULTIPROC_METHOD=spawn \
+    VLLM_ASCEND_ENABLE_VNPU=1
+
 # Install modelscope (for fast download) and ray (for multinode)
 RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
     python3 -m pip cache purge

Dockerfile.a3.openEuler

@@ -20,8 +20,10 @@ FROM quay.io/ascend/cann:8.3.rc2-a3-openeuler24.03-py3.11
 ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
 ARG COMPILE_CUSTOM_KERNELS=1
 ARG MOONCAKE_TAG="v0.3.7.post2"
+ARG SOC_VERSION="ascend910_9391"
 
-ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS} \
+    SOC_VERSION=${SOC_VERSION}
 
 RUN pip config set global.index-url ${PIP_INDEX_URL}
 
@@ -61,9 +63,15 @@ RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
     export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
+    cd /vllm-workspace/vllm-ascend/csrc/vnpu_offload && \
+    make install && make clean && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
     python3 -m pip cache purge
 
+ENV VLLM_ASCEND_ENABLE_NZ=0 \
+    VLLM_WORKER_MULTIPROC_METHOD=spawn \
+    VLLM_ASCEND_ENABLE_VNPU=1
+
 # Install modelscope (for fast download) and ray (for multinode)
 RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
     python3 -m pip cache purge

Dockerfile.openEuler

@@ -62,9 +62,15 @@ RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
     source /usr/local/Ascend/nnal/atb/set_env.sh && \
     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
     export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
+    cd /vllm-workspace/vllm-ascend/csrc/vnpu_offload && \
+    make install && make clean && \
     python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
     python3 -m pip cache purge
 
+ENV VLLM_ASCEND_ENABLE_NZ=0 \
+    VLLM_WORKER_MULTIPROC_METHOD=spawn \
+    VLLM_ASCEND_ENABLE_VNPU=1
+
 # Install modelscope (for fast download) and ray (for multinode)
 RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
     python3 -m pip cache purge

Dockerfile_backup/Dockerfile.310p

@@ -0,0 +1,61 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+#
+
+FROM quay.io/ascend/cann:8.3.rc2-310p-ubuntu22.04-py3.11
+
+ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
+ARG COMPILE_CUSTOM_KERNELS=1
+
+# Define environments
+ENV DEBIAN_FRONTEND=noninteractive
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+
+RUN apt-get update -y && \
+    apt-get install -y python3-pip git vim wget net-tools gcc g++ cmake libnuma-dev && \
+    rm -rf /var/cache/apt/* && \
+    rm -rf /var/lib/apt/lists/*
+
+WORKDIR /workspace
+
+COPY . /vllm-workspace/vllm-ascend/
+
+RUN pip config set global.index-url ${PIP_INDEX_URL}
+
+# Install vLLM
+ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
+ARG VLLM_TAG=v0.11.0
+RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /vllm-workspace/vllm
+# In x86, triton will be installed by vllm. But in Ascend, triton doesn't work correctly. we need to uninstall it.
+RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -v -e /vllm-workspace/vllm/[audio] --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip uninstall -y triton && \
+    python3 -m pip cache purge
+
+# Install vllm-ascend
+# Append `libascend_hal.so` path (devlib) to LD_LIBRARY_PATH
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    source /usr/local/Ascend/nnal/atb/set_env.sh && \
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
+    export SOC_VERSION=ASCEND310P3 && \
+    python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip cache purge
+
+# Install modelscope (for fast download) and ray (for multinode)
+RUN python3 -m pip install modelscope 'ray>=2.47.1' 'protobuf>3.20.0' && \
+    python3 -m pip cache purge
+
+CMD ["/bin/bash"]

Dockerfile_backup/Dockerfile.310p.openEuler

@@ -0,0 +1,59 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+#
+
+FROM quay.io/ascend/cann:8.3.rc2-310p-openeuler24.03-py3.11
+
+ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
+ARG COMPILE_CUSTOM_KERNELS=1
+
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+
+RUN yum update -y && \
+    yum install -y python3-pip git vim wget net-tools gcc gcc-c++ make cmake numactl-devel && \
+    rm -rf /var/cache/yum
+
+RUN pip config set global.index-url ${PIP_INDEX_URL}
+
+WORKDIR /workspace
+
+COPY . /vllm-workspace/vllm-ascend/
+
+# Install vLLM
+ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
+ARG VLLM_TAG=v0.11.0
+
+RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /vllm-workspace/vllm
+# In x86, triton will be installed by vllm. But in Ascend, triton doesn't work correctly. we need to uninstall it.
+RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -e /vllm-workspace/vllm/[audio] --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip uninstall -y triton && \
+    python3 -m pip cache purge
+
+# Install vllm-ascend
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    source /usr/local/Ascend/nnal/atb/set_env.sh && \
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
+    export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
+    export SOC_VERSION=ASCEND310P3 && \
+    python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip cache purge
+
+# Install modelscope (for fast download) and ray (for multinode)
+RUN python3 -m pip install modelscope 'ray>=2.47.1' 'protobuf>3.20.0' && \
+    python3 -m pip cache purge
+
+CMD ["/bin/bash"]

Dockerfile_backup/Dockerfile.a3

@@ -0,0 +1,68 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+#
+
+FROM quay.io/ascend/cann:8.3.rc2-a3-ubuntu22.04-py3.11
+
+ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
+ARG COMPILE_CUSTOM_KERNELS=1
+ARG MOONCAKE_TAG=v0.3.7.post2
+
+COPY . /vllm-workspace/vllm-ascend/
+# Define environments
+ENV DEBIAN_FRONTEND=noninteractive
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+
+RUN pip config set global.index-url ${PIP_INDEX_URL}
+
+WORKDIR /workspace
+
+# Install Mooncake dependencies
+RUN apt-get update -y && \
+    apt-get install -y git vim wget net-tools gcc g++ cmake libnuma-dev && \
+    git clone --depth 1 --branch ${MOONCAKE_TAG} https://github.com/kvcache-ai/Mooncake /vllm-workspace/Mooncake && \
+    cp /vllm-workspace/vllm-ascend/tools/mooncake_installer.sh /vllm-workspace/Mooncake/ && \
+    cd /vllm-workspace/Mooncake && bash mooncake_installer.sh -y && \
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/lib64 && \
+    mkdir -p build && cd build && cmake .. -DUSE_ASCEND_DIRECT=ON && \
+    make -j$(nproc) && make install && \
+    rm -fr /vllm-workspace/Mooncake/build && \
+    rm -rf /var/cache/apt/* && \
+    rm -rf /var/lib/apt/lists/*
+
+# Install vLLM
+ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
+ARG VLLM_TAG=v0.11.0
+RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /vllm-workspace/vllm
+# In x86, triton will be installed by vllm. But in Ascend, triton doesn't work correctly. we need to uninstall it.
+RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -v -e /vllm-workspace/vllm/[audio] --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip uninstall -y triton && \
+    python3 -m pip cache purge
+
+# Install vllm-ascend
+# Append `libascend_hal.so` path (devlib) to LD_LIBRARY_PATH
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    source /usr/local/Ascend/nnal/atb/set_env.sh && \
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
+    python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip cache purge
+
+# Install modelscope (for fast download) and ray (for multinode)
+RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
+    python3 -m pip cache purge
+
+CMD ["/bin/bash"]

Dockerfile_backup/Dockerfile.a3.openEuler

@@ -0,0 +1,71 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+#
+
+FROM quay.io/ascend/cann:8.3.rc2-a3-openeuler24.03-py3.11
+
+ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
+ARG COMPILE_CUSTOM_KERNELS=1
+ARG MOONCAKE_TAG="v0.3.7.post2"
+
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+
+RUN pip config set global.index-url ${PIP_INDEX_URL}
+
+WORKDIR /workspace
+
+COPY . /vllm-workspace/vllm-ascend/
+
+SHELL ["/bin/bash", "-c"]
+
+RUN yum update -y && \
+    yum install -y git vim wget net-tools gcc gcc-c++ make cmake numactl-devel && \
+    git clone --depth 1 --branch ${MOONCAKE_TAG} https://github.com/kvcache-ai/Mooncake /vllm-workspace/Mooncake && \
+    cp /vllm-workspace/vllm-ascend/tools/mooncake_installer.sh /vllm-workspace/Mooncake/ && \
+    ARCH=$(uname -m) && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    export LD_LIBRARY_PATH=/usr/local/Ascend/ascend-toolkit/latest/${ARCH}-linux/devlib:/usr/local/Ascend/ascend-toolkit/latest/${ARCH}-linux/lib64:$LD_LIBRARY_PATH && \
+    export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/${ARCH}-openEuler-linux && \
+    cd /vllm-workspace/Mooncake && \
+    bash mooncake_installer.sh -y && \
+    mkdir -p build && cd build && cmake .. -DUSE_ASCEND_DIRECT=ON && \
+    make -j$(nproc) && make install && \
+    rm -fr /vllm-workspace/Mooncake/build && \
+    rm -rf /var/cache/yum/*
+
+# Install vLLM
+ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
+ARG VLLM_TAG=v0.11.0
+RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /vllm-workspace/vllm
+# In x86, triton will be installed by vllm. But in Ascend, triton doesn't work correctly. we need to uninstall it.
+RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -e /vllm-workspace/vllm/[audio] --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip uninstall -y triton && \
+    python3 -m pip cache purge
+
+# Install vllm-ascend
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    source /usr/local/Ascend/nnal/atb/set_env.sh && \
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
+    export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
+    python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip cache purge
+
+# Install modelscope (for fast download) and ray (for multinode)
+RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
+    python3 -m pip cache purge
+
+CMD ["/bin/bash"]

Dockerfile_backup/Dockerfile.openEuler

@@ -0,0 +1,72 @@
+#
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# This file is a part of the vllm-ascend project.
+#
+
+FROM quay.io/ascend/cann:8.3.rc2-910b-openeuler24.03-py3.11
+
+ARG PIP_INDEX_URL="https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple"
+ARG COMPILE_CUSTOM_KERNELS=1
+ARG MOONCAKE_TAG="v0.3.7.post2"
+
+ENV COMPILE_CUSTOM_KERNELS=${COMPILE_CUSTOM_KERNELS}
+
+
+RUN pip config set global.index-url ${PIP_INDEX_URL}
+
+WORKDIR /workspace
+
+COPY . /vllm-workspace/vllm-ascend/
+
+SHELL ["/bin/bash", "-c"]
+
+RUN yum update -y && \
+    yum install -y git vim wget net-tools gcc gcc-c++ make cmake numactl-devel && \
+    git clone --depth 1 --branch ${MOONCAKE_TAG} https://github.com/kvcache-ai/Mooncake /vllm-workspace/Mooncake && \
+    cp /vllm-workspace/vllm-ascend/tools/mooncake_installer.sh /vllm-workspace/Mooncake/ && \
+    ARCH=$(uname -m) && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    export LD_LIBRARY_PATH=/usr/local/Ascend/ascend-toolkit/latest/${ARCH}-linux/devlib:/usr/local/Ascend/ascend-toolkit/latest/${ARCH}-linux/lib64:$LD_LIBRARY_PATH && \
+    export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/${ARCH}-openEuler-linux && \
+    cd /vllm-workspace/Mooncake && \
+    bash mooncake_installer.sh -y && \
+    mkdir -p build && cd build && cmake .. -DUSE_ASCEND_DIRECT=ON && \
+    make -j$(nproc) && make install && \
+    rm -fr /vllm-workspace/Mooncake/build && \
+    rm -rf /var/cache/yum/*
+
+# Install vLLM
+ARG VLLM_REPO=https://github.com/vllm-project/vllm.git
+ARG VLLM_TAG=v0.11.0
+RUN git clone --depth 1 $VLLM_REPO --branch $VLLM_TAG /vllm-workspace/vllm
+# In x86, triton will be installed by vllm. But in Ascend, triton doesn't work correctly. we need to uninstall it.
+RUN VLLM_TARGET_DEVICE="empty" python3 -m pip install -e /vllm-workspace/vllm/[audio] --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip uninstall -y triton && \
+    python3 -m pip cache purge
+
+# Install vllm-ascend
+RUN export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
+    source /usr/local/Ascend/nnal/atb/set_env.sh && \
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
+    export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/include/c++/12:/usr/include/c++/12/`uname -i`-openEuler-linux && \
+    python3 -m pip install -v -e /vllm-workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/ && \
+    python3 -m pip cache purge
+
+# Install modelscope (for fast download) and ray (for multinode)
+RUN python3 -m pip install modelscope 'ray>=2.47.1,<=2.48.0' 'protobuf>3.20.0' && \
+    python3 -m pip cache purge
+
+CMD ["/bin/bash"]

README.md

@@ -1,18 +1,10 @@
-# vLLM-Ascend Multi-LLM Serving Support
+# XC-LLM: A Specially Optimized LLM Inference Engine for ModelHub XC
 
 ## Overview
 
-This repository is a modified version of [vLLM-Ascend](https://github.com/vllm-project/vllm-ascend) designed to enable multiple large language models (LLMs) to share one Ascend NPU.
+The project is optimized based on the popular LLM inference project vLLM. The current version supports Ascend NPU (910B3 & 910B4). 
 
-The key feature of this project is efficient memory coordination, enabling multiple vLLM instances share and dynamically hold Ascend NPU's physical memory.
+One of the key features of this project is efficient memory coordination, enabling multiple vLLM instances share and dynamically hold Ascend NPU's physical memory. When an instance is idle, model parameters are offloaded to host memory. Upon a new inference request, the model parameters are quickly restored to the NPU’s memory (if not exist), without the need to initialize the engine and load the model from scratch. As a result, from the application’s perspective, multiple LLM inference engines can run on the NPU even when their total memory requirements exceed the physical memory limit. This technique is referred to as `InfiniVRAM`.
-When an instance is idle, model parameters are offloaded to host memory.
-Upon a new inference request, the model parameters are quickly restored to the NPU’s memory (if not exist), without the need to init the engine and load the model from scratch. (For Qwen3-8B, It only causes 0.3s of additional latency to TTFT on a real restore.)
-
-
-## Features
-
-- **Shared NPU Usage**: Multiple vLLM instances can access the same Ascend NPU, allowing for multi-LLM serving of different LLMs.
-- **Fast Memory Restore**: We decouple the virtual and physical memory allcation. Physical NPU memory is allocated and exported and shared to other LLM engines. LLM engines can restore quickly without reinitialize and memory allocation
 
 
 ## Installation
@@ -33,19 +25,14 @@ docker build -t vllm-ascend-multi-llm:latest -f ./Dockerfile .
 0. To share NPU, processes coordinate via shm, so you need to set all containers with `ipc=host`.
 1. Start a daemon process in a standalone container, by running `vllm_vnpu_daemon` installed inside the image.
 2. Start LLM services with this image, following the official usage instructions.
+3. Due to the limited stream resource of Ascend NPU, you may need to restrict graph capture sizes or disable ACLgraph by setting `--enforce-eager`, especially when launching multiple LLMs. Refer to the [link](https://docs.vllm.ai/projects/ascend/en/latest/faqs.html#how-to-troubleshoot-and-resolve-size-capture-failures-resulting-from-stream-resource-exhaustion-and-what-are-the-underlying-causes).
+
+### Environment Variables
+- `VNPU_RESERVED_VRAM_SIZE_GB`: The amonut of reserved GPU memory for other miscellaneous memory. Only needs to be set for `vllm_vnpu_daemon`. Try increasing the variable if you launch multiple LLM services and encounter OOM. Default: `8`.
+- `VLLM_VNPU_SHM_NAME`: The name of the shm file. Needs to be set for all containers of the shared vNPU group. Default: `/vllm_acl_vnpu_offload_shm`.
 
 
 ## Limitations
 
-- This project only support share a single NPU currently. This is also limited by the fact that HCCL cannot be shared. We haven't figure out how to bypass HCCL. *Help wanted*.
+- Restricted by the fact that HCCL cannot be shared, deploying more than one model with multi-GPU (e.g., TP) is not feasible currently.
 - The prefix cache will be reset when the LLM is restored, since we just simply discard the KV cache when the LLM is offloaded.
-
-
-## Roadmap
-- [ ] Space-sharing.
-- [ ] ...
-
-
-## License
-
-Apache License 2.0, as found in the [LICENSE](./LICENSE) file.

csrc/camem_allocator.cpp

@@ -19,7 +19,8 @@
 #include <string>
 #include <atomic>
 
-#include "idle_offload/shm_worker.h"
+#include "vnpu_offload/shm_worker.h"
+#include "vnpu_offload/npu_helper.h"
 
 extern "C" {
 
@@ -311,7 +312,8 @@ my_malloc_offload(ssize_t size, int device, aclrtStream stream) {
       (aclrtDrvMemHandle*)malloc(sizeof(aclrtDrvMemHandle));
 
   if (!g_python_malloc_callback) {
-    throw std::runtime_error("my_malloc ERROR: g_python_malloc_callback not set." +
+    throw std::runtime_error(
+        "my_malloc ERROR: g_python_malloc_callback not set." +
         std::string(" ") + __FILE__ + ":" + std::to_string(__LINE__));
   }
 
@@ -345,8 +347,9 @@ __attribute__((visibility("default"))) void
 my_free_offload(void *ptr, ssize_t size, int device, aclrtStream stream) {
   // get memory handle from the pointer
   if (!g_python_free_callback) {
-    throw std::runtime_error("aclrtDrvMemHandle ERROR: g_python_malloc_callback not set." +
+    throw std::runtime_error(
-                            std::string(" ") + __FILE__ + ":" + std::to_string(__LINE__));
+        "my_free ERROR: g_python_malloc_callback not set." + std::string(" ") +
+        __FILE__ + ":" + std::to_string(__LINE__));
   }
 
   // Acquire GIL (not in stable ABI officially, but often works)
@@ -474,8 +477,10 @@ static PyObject* python_create_and_map(PyObject* self, PyObject* args) {
 static PyObject* py_init_module_offload(PyObject* self, PyObject* args) {
   PyObject* malloc_callback = nullptr;
   PyObject* free_callback = nullptr;
+  unsigned long long device = 0;
 
-  if (!PyArg_ParseTuple(args, "OO", &malloc_callback, &free_callback)) {
+  if (!PyArg_ParseTuple(args, "OOK", &malloc_callback, &free_callback,
+                        &device)) {
     return nullptr;
   }
 
@@ -497,7 +502,13 @@ static PyObject* py_init_module_offload(PyObject* self, PyObject* args) {
   }
   g_initialized.store(true);
 
-  shm_worker = new ShmWorker();
+  std::vector<int> gpu_ids = get_npu_ids();
+  if (device >= gpu_ids.size()) {
+    throw std::runtime_error("Invalid device id: " + std::to_string(device) +
+                             " " + __FILE__ + ":" + std::to_string(__LINE__));
+  }
+  int gpu_id = gpu_ids[device];
+
   // get pid
   aclError error_code;
   int32_t pid;
@@ -508,11 +519,12 @@ static PyObject* py_init_module_offload(PyObject* self, PyObject* args) {
         std::to_string(error_code) + " " + __FILE__ + ":" +
         std::to_string(__LINE__));
   }
+
+  shm_worker = new ShmWorker();
   uint64_t shareable_handle;
-  shm_worker->register_worker(pid, &shareable_handle, &g_size);
+  shm_worker->register_worker(pid, gpu_id, &shareable_handle, &g_size);
 
   // import shareable handle
-  uint32_t device = 0;
   aclrtDrvMemHandle memHandle;
   error_code =
       aclrtMemImportFromShareableHandle(shareable_handle, device, &memHandle);
@@ -570,9 +582,16 @@ static PyObject* python_get_mem_info_offload(PyObject* self, PyObject* args) {
   return tuple;
 }
 
-static PyObject* python_lock_gpu_offload(PyObject* self, PyObject* args) {
+static PyObject* python_try_lock_gpu_offload(PyObject* self, PyObject* args) {
-  bool prev_is_self = shm_worker->lock_gpu();
+  bool prev_is_self = false;
-  return PyBool_FromLong(prev_is_self);
+  bool success = shm_worker->try_lock_gpu(prev_is_self);
+  PyObject* tuple = PyTuple_New(2);
+  if (!tuple) {  
+    return nullptr;
+  }
+  PyTuple_SetItem(tuple, 0, PyBool_FromLong(success));
+  PyTuple_SetItem(tuple, 1, PyBool_FromLong(prev_is_self));
+  return tuple;
 }
 
 static PyObject* python_unlock_gpu_offload(PyObject* self, PyObject* args) {
@@ -597,7 +616,7 @@ static PyMethodDef module_methods[] = {
      "Unmap and release memory on the device."},
     {"python_get_mem_info_offload", (PyCFunction)python_get_mem_info_offload,
      METH_NOARGS, "Get mem info in the reserved pool."},
-    {"python_lock_gpu_offload", (PyCFunction)python_lock_gpu_offload,
+    {"python_try_lock_gpu_offload", (PyCFunction)python_try_lock_gpu_offload,
      METH_NOARGS, "Lock GPU."},
     {"python_unlock_gpu_offload", (PyCFunction)python_unlock_gpu_offload,
      METH_NOARGS, "Unlock GPU."},

csrc/idle_offload/offload_daemon.cpp

@@ -1,198 +0,0 @@
-#include <iostream>
-#include <sys/types.h>
-
-#include <sys/mman.h>
-#include <sys/stat.h>
-#include <fcntl.h>
-#include <unistd.h>
-#include <string.h>
-#include <vector>
-#include <atomic>
-#include <signal.h>
-
-#include "acl/acl.h"
-
-#include "shm_manager.h"
-#include "spdlog/spdlog.h"
-
-
-static constexpr size_t reserved_mem_size = 8ul * 1024 * 1024 * 1024; // 8GB
-
-static ShmManager *shm_manager = nullptr;
-
-void handle_signal(int sig) {
-  if (shm_manager) {
-    shm_manager->stop_busy_loop();
-  }
-}
-
-void install_signal_handlers() {
-    struct sigaction sa{};
-    sa.sa_handler = handle_signal;
-    sigemptyset(&sa.sa_mask);
-    sa.sa_flags = 0;
-
-    sigaction(SIGINT,  &sa, nullptr);
-    sigaction(SIGTERM, &sa, nullptr);
-    sigaction(SIGHUP,  &sa, nullptr);
-}
-
-void ensure_context(unsigned long long device) {
-  aclrtContext pctx;
-  aclrtGetCurrentContext(&pctx);
-  if (!pctx) {
-    // Ensure device context.
-    aclrtCreateContext(&pctx, device);
-    aclrtSetCurrentContext(pctx);
-  }
-}
-
-void init_acl() {
-  int32_t deviceId=0;
-  // aclrtStream stream;
-  bool g_isDevice;
-
-  aclError ret = aclrtSetDevice(deviceId);
-  if (ret != ACL_ERROR_NONE) {
-    throw std::runtime_error("aclrtSetDevice failed with acl error code: " +
-                            std::to_string(ret) + " " + __FILE__ + ":" + std::to_string(__LINE__));
-  }
-}
-
-void reset_pids(const std::vector<int32_t> &pids, uint64_t shareable_handle) {
-  int cnt = pids.size();
-  if (cnt <= 0) {
-    return;
-  }
-
-  int32_t pids_data[cnt];
-  memcpy(pids_data, pids.data(), cnt * sizeof(int32_t));
-
-  aclError error_code =
-      aclrtMemSetPidToShareableHandle(shareable_handle, pids_data, cnt);
-  if (error_code != 0) {
-    spdlog::error("aclrtMemSetPidToShareableHandle failed, error_code: {}",
-                  error_code);
-    throw std::runtime_error("aclrtMemSetPidToShareableHandle failed");
-  } else {
-    spdlog::info("aclrtMemSetPidToShareableHandle succeeded, num_pids: {}",
-                 cnt);
-  }
-}
-
-void start_daemon() {
-  init_acl();
-
-  aclError error_code;
-  size_t free_mem = 0, total = 0;
-  error_code = aclrtGetMemInfo(ACL_HBM_MEM, &free_mem, &total);
-  if (error_code != 0) {
-    spdlog::error("aclrtGetMemInfo failed, error_code: {}", error_code);
-    throw std::runtime_error("aclrtGetMemInfo failed");
-  } else {
-    spdlog::info("aclrtGetMemInfo succeeded, free_mem: {}, total: {}", free_mem,
-                 total);
-  }
-
-  uint32_t device = 0;
-  aclrtPhysicalMemProp prop = {};
-  prop.handleType = ACL_MEM_HANDLE_TYPE_NONE;
-  prop.allocationType = ACL_MEM_ALLOCATION_TYPE_PINNED;
-  prop.memAttr = ACL_HBM_MEM_HUGE;
-  prop.location.id = device;
-  prop.location.type = ACL_MEM_LOCATION_TYPE_DEVICE;
-  prop.reserve = 0;
-
-  size_t granularity;
-  error_code = aclrtMemGetAllocationGranularity(
-      &prop, ACL_RT_MEM_ALLOC_GRANULARITY_MINIMUM, &granularity);
-  if (error_code != 0) {
-    spdlog::error("aclrtMemGetAllocationGranularity failed, error_code: {}", error_code);
-    throw std::runtime_error("aclrtMemGetAllocationGranularity failed");
-  } else {
-    spdlog::info("aclrtMemGetAllocationGranularity succeeded, granularity: {}",
-                 granularity);
-  }
-  if (free_mem < reserved_mem_size) {
-    spdlog::error("Not enough free memory to reserve: {}, free_mem: {}",
-                  reserved_mem_size, free_mem);
-    throw std::runtime_error("Not enough free memory to reserve");
-  }
-  size_t g_size = free_mem - reserved_mem_size;
-  g_size = (g_size / granularity) * granularity;
-
-  // allocate physical memory
-  aclrtDrvMemHandle mem_handle;
-  error_code = aclrtMallocPhysical(&mem_handle, g_size, &prop, 0);
-  if (error_code != 0) {
-    spdlog::error("aclrtMallocPhysical failed, error_code: {}", error_code);
-    throw std::runtime_error("aclrtMallocPhysical failed");
-  } else {
-    spdlog::info("aclrtMallocPhysical succeeded, size: {}", g_size);
-  }
-
-  // // reserve address
-  // void *vmem_addr = nullptr;
-  // error_code = aclrtReserveMemAddress(&vmem_addr, g_size, 0, nullptr, 0);
-  // if (error_code != 0) {
-  //   spdlog::error("aclrtReserveMemAddress failed, error_code: {}", error_code);
-  //   throw std::runtime_error("aclrtReserveMemAddress failed");
-  // } else {
-  //   spdlog::info("aclrtReserveMemAddress succeeded, vmem_addr: {}", vmem_addr);
-  // }
-
-  // // map
-  // error_code = aclrtMapMem(vmem_addr, g_size, 0, mem_handle, 0);
-  // if (error_code != 0) {
-  //   spdlog::error("aclrtMapMem failed, error_code: {}", error_code);
-  //   throw std::runtime_error("aclrtMapMem failed");
-  // } else {
-  //   spdlog::info("aclrtMapMem succeeded, vmem_addr: {}", vmem_addr);
-  // }
-
-  // export
-  uint64_t shareable_handle;
-  error_code = aclrtMemExportToShareableHandle(
-      mem_handle, ACL_MEM_HANDLE_TYPE_NONE, ACL_RT_VMM_EXPORT_FLAG_DEFAULT,
-      &shareable_handle);
-  if (error_code != 0) {
-    spdlog::error("aclrtMemExportToShareableHandle failed, error_code: {}",
-                  error_code);
-    throw std::runtime_error("aclrtMemExportToShareableHandle failed");
-  } else {
-    spdlog::info(
-        "aclrtMemExportToShareableHandle succeeded, shareable_handle: {}",
-        shareable_handle);
-  }
-
-  // shm
-  shm_manager = new ShmManager();
-  shm_manager->set_gpu_info(g_size, shareable_handle);
-  shm_manager->register_callback_on_worker_change(
-      [&](const std::vector<int32_t> &pids) {
-        reset_pids(pids, shareable_handle);
-      });
-
-  // start busy loop
-  shm_manager->run_busy_loop();
-
-  // stopped by signal
-  delete shm_manager;
-  shm_manager = nullptr;
-
-  // free physical memory
-  error_code = aclrtFreePhysical(mem_handle);
-  if (error_code != 0) {
-    spdlog::error("aclrtFreePhysical failed, error_code: {}", error_code);
-    throw std::runtime_error("aclrtFreePhysical failed");
-  }
-}
-
-
-int main() {
-  install_signal_handlers();
-
-  start_daemon();
-
-  return 0;
-}

csrc/vnpu_offload/Makefile

@@ -1,12 +1,12 @@
 CXX := g++
 TARGET := vllm_vnpu_daemon
-SRCS := offload_daemon.cpp shm_manager.cpp
+SRCS := vnpu_daemon.cpp shm_manager.cpp
 
 ASCEND_HOME := /usr/local/Ascend/ascend-toolkit/latest
 INCLUDES := -I$(ASCEND_HOME)/include -Iinclude
 LIBS := -L$(ASCEND_HOME)/lib64 -lascendcl
 
-CXXFLAGS := $(INCLUDES)
+CXXFLAGS := $(INCLUDES) -O2
 LDFLAGS := $(LIBS)
 
 PREFIX ?= /usr/local