install_enflame_cn.md

July 12, 2026 · View on GitHub

[中文版|English]

💫 Enflame（燧原）enflame (Triton 3.6)

对应的 Triton 版本为 3.6，基于 x64 平台
可用于 GCU300 (S60), GCU400 (L300/L600)

1. 构建及运行环境

1.1 使用镜像 (Triton 3.6, GCU300/GCU400)

如果网络环境畅通，不必执行后续步骤 1.x，依赖库会在构建时自动拉取。

# Plan A: docker pull (31.6GB)
IMAGE=harbor.baai.ac.cn/flagtree/flagtree-enflame3.6-py312-torch2.10.0-ubuntu24.04:202606-1.9.17-base
docker pull ${IMAGE}
# Plan B: docker load (6.4GB)
IMAGE=flagtree-enflame3.6-py312-torch2.10.0-ubuntu24.04:202606-1.9.17-base
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/flagtree-enflame3.6-py312-torch2.10.0-ubuntu24.04.202606-1.9.17-base.tar.gz
docker load -i flagtree-enflame3.6-py312-torch2.10.0-ubuntu24.04.202606-1.9.17-base.tar.gz

cat /sys/module/enflame/version
    # if version < 1.9.17, terminate the processes using GCU, and execute the following commands on the host:
    # wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/TopsRider_Triton_gcu-3.6.0_1.0.20260610.cc.1.9.17_deb_amd64.run  # 1.8GB
    # bash TopsRider_Triton_gcu-3.6.0_1.0.20260610.cc.1.9.17_deb_amd64.run --driver -y
efsmi
CONTAINER=flagtree-dev-xxx
docker run -dit \
    --privileged \
    -v /etc/localtime:/etc/localtime:ro \
    -v /home:/home \
    -w /root --name ${CONTAINER} ${IMAGE} bash
docker exec -it ${CONTAINER} /bin/bash

1.2 手动下载 FlagTree 依赖库

mkdir -p ~/.flagtree/enflame; cd ~/.flagtree/enflame
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/enflame-llvm23-fc83c68-gcc9-x64_v0.4.0.tar.gz
tar zxvf enflame-llvm23-fc83c68-gcc9-x64_v0.4.0.tar.gz

1.3 手动下载 Triton 依赖库

镜像中已下载安装 Triton 依赖库。如果无需从源码构建 FlagTree 或 Triton，那么无需下载 Triton 依赖库。

cd ${YOUR_CODE_DIR}/FlagTree
# For Triton 3.6 (x64)
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/build-deps-triton_3.6.x-linux-x64.tar.gz
sh python/scripts/unpack_triton_build_deps.sh ./build-deps-triton_3.6.x-linux-x64.tar.gz

执行完上述脚本后，原有的 ~/.triton 目录将被重命名，新的 ~/.triton 目录会被创建并存放预下载包。注意执行脚本过程中会提示手动确认。

2. 安装命令

2.1 免源码安装

# Note: First install PyTorch, then execute the following commands
python3 -m pip uninstall -y triton --break-system-packages  # Repeat the cmd until fully uninstalled
RES="--index-url=https://resource.flagos.net/repository/flagos-pypi-hosted/simple"
python3.12 -m pip install flagtree===0.6.0+enflame3.6 --break-system-packages $RES

安装 flagtree 后，可通过下列命令查看：

python3 -m pip show flagtree

2.2 从源码构建

cd ${YOUR_CODE_DIR}/FlagTree
git checkout main
export FLAGTREE_BACKEND=enflame
MAX_JOBS=8 python3 -m pip install . --no-build-isolation -v --break-system-packages

3. 测试验证

参考 Tests of enflame3.6 backend

💫 Enflame（燧原）enflame (Triton 3.5)

对应的 Triton 版本为 3.5，基于 x64 平台
可用于 GCU300 (S60), GCU400 (L300/L600)

1. 构建及运行环境

1.1 使用镜像 (Triton 3.5, GCU300/GCU400)

如果网络环境畅通，不必执行后续步骤 1.x，依赖库会在构建时自动拉取。

# Plan A: docker pull (13.3GB)
IMAGE=harbor.baai.ac.cn/flagtree/flagtree-enflame3.5-py312-torch2.9.1-ubuntu24.04:202603
docker pull ${IMAGE}
# Plan B: docker load (2.8GB)
IMAGE=flagtree-enflame3.5-py312-torch2.9.1-ubuntu24.04:202603
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/flagtree-enflame3.5-py312-torch2.9.1-ubuntu24.04.202603.tar.gz
docker load -i flagtree-enflame3.5-py312-torch2.9.1-ubuntu24.04.202603.tar.gz

CONTAINER=flagtree-dev-xxx
docker run -dit \
    --privileged \
    -v /etc/localtime:/etc/localtime:ro \
    -v /home:/home \
    -w /root --name ${CONTAINER} ${IMAGE} bash
docker cp ${CONTAINER}:/enflame enflame    # Will create ./enflame dir
bash enflame/driver/enflame-x86_64-gcc-1.7.2.14-20260302150833.run
efsmi
docker stop ${CONTAINER}
docker start ${CONTAINER}
docker exec -it ${CONTAINER} /bin/bash

1.2 手动下载 FlagTree 依赖库

mkdir -p ~/.flagtree/enflame; cd ~/.flagtree/enflame
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/enflame-llvm22-189e06b-gcc9-x64_v0.4.0.tar.gz
tar zxvf enflame-llvm22-189e06b-gcc9-x64_v0.4.0.tar.gz

1.3 手动下载 Triton 依赖库

镜像中已下载安装 Triton 依赖库。如果无需从源码构建 FlagTree 或 Triton，那么无需下载 Triton 依赖库。

cd ${YOUR_CODE_DIR}/FlagTree
# For Triton 3.5 (x64)
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/build-deps-triton_3.5.x-linux-x64.tar.gz
sh python/scripts/unpack_triton_build_deps.sh ./build-deps-triton_3.5.x-linux-x64.tar.gz

执行完上述脚本后，原有的 ~/.triton 目录将被重命名，新的 ~/.triton 目录会被创建并存放预下载包。注意执行脚本过程中会提示手动确认。

2. 安装命令

2.1 免源码安装

# Note: First install PyTorch, then execute the following commands
python3 -m pip uninstall -y triton --break-system-packages  # Repeat the cmd until fully uninstalled
RES="--index-url=https://resource.flagos.net/repository/flagos-pypi-hosted/simple"
python3.12 -m pip install flagtree===0.5.0+enflame3.5 --break-system-packages $RES

安装 flagtree 后，可通过下列命令查看：

python3 -m pip show flagtree

2.2 从源码构建

cd ${YOUR_CODE_DIR}/FlagTree
git checkout -b triton_v3.5.x origin/triton_v3.5.x
export FLAGTREE_BACKEND=enflame
MAX_JOBS=8 python3 -m pip install . --no-build-isolation -v --break-system-packages

3. 测试验证

参考 Tests of enflame3.5 backend

💫 Enflame（燧原）enflame (Triton 3.3)

对应的 Triton 版本为 3.3，基于 x64 平台
可用于 GCU300 (S60)

1. 构建及运行环境

1.1 使用镜像 (Triton 3.3, GCU300)

如果网络环境畅通，不必执行后续步骤 1.x，依赖库会在构建时自动拉取。

# Plan A: docker pull (12.5GB)
IMAGE=harbor.baai.ac.cn/flagtree/flagtree-enflame3.3-py310-torch2.7.0-ubuntu22.04:202603
docker pull ${IMAGE}
# Plan B: docker load (5.7GB)
IMAGE=flagtree-enflame3.3-py310-torch2.7.0-ubuntu22.04:202603
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/flagtree-hcu-py310-torch2.9.0-ubuntu22.04.202603.tar.gz
docker load -i flagtree-hcu-py310-torch2.9.0-ubuntu22.04.202603.tar.gz

CONTAINER=flagtree-dev-xxx
docker run -dit \
    --privileged \
    -v /etc/localtime:/etc/localtime:ro \
    -v /home:/home \
    -w /root --name ${CONTAINER} ${IMAGE} bash
docker cp ${CONTAINER}:/enflame enflame    # Will create ./enflame dir
bash enflame/driver/enflame-x86_64-gcc-1.6.3.12-20251115104629.run
efsmi
docker stop ${CONTAINER}
docker start ${CONTAINER}
docker exec -it ${CONTAINER} /bin/bash

1.2 手动下载 FlagTree 依赖库

mkdir -p ~/.flagtree/enflame; cd ~/.flagtree/enflame
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/enflame-llvm21-d752c5b-gcc9-x64_v0.3.0.tar.gz
tar zxvf enflame-llvm21-d752c5b-gcc9-x64_v0.3.0.tar.gz

1.3 手动下载 Triton 依赖库

镜像中已下载安装 Triton 依赖库。如果无需从源码构建 FlagTree 或 Triton，那么无需下载 Triton 依赖库。

cd ${YOUR_CODE_DIR}/FlagTree
# For Triton 3.3 (x64)
wget https://baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/build-deps-triton_3.3.x-linux-x64.tar.gz
sh python/scripts/unpack_triton_build_deps.sh ./build-deps-triton_3.3.x-linux-x64.tar.gz

执行完上述脚本后，原有的 ~/.triton 目录将被重命名，新的 ~/.triton 目录会被创建并存放预下载包。注意执行脚本过程中会提示手动确认。

2. 安装命令

2.1 免源码安装

# Note: First install PyTorch, then execute the following commands
python3 -m pip uninstall -y triton  # Repeat the cmd until fully uninstalled
RES="--index-url=https://resource.flagos.net/repository/flagos-pypi-hosted/simple"
python3.10 -m pip install flagtree===0.4.0+enflame3.3 $RES

安装 flagtree 后，可通过下列命令查看：

python3 -m pip show flagtree

2.2 从源码构建

cd ${YOUR_CODE_DIR}/FlagTree/python
git checkout -b triton_v3.3.x origin/triton_v3.3.x
export FLAGTREE_BACKEND=enflame
MAX_JOBS=8 python3 -m pip install . --no-build-isolation -v

3. 测试验证

参考 Tests of enflame3.3 backend