Skip to content

README_cn.md

Latest commit

 

History

History
158 lines (144 loc) · 7.32 KB

README_cn.md

File metadata and controls

158 lines (144 loc) · 7.32 KB

English

FlagTree

FlagTree 是面向多种 AI 芯片的开源、统一编译器。FlagTree 致力于打造多元 AI 芯片编译器及相关工具平台，发展和壮大 Triton 上下游生态。项目当前处于初期，目标是兼容现有适配方案，统一代码仓库，快速实现单仓库多后端支持。对于上游模型用户，提供多后端的统一编译能力；对于下游芯片厂商，提供 Triton 生态接入范例。

从源代码安装

安装依赖（注意使用正确的 python3.x 执行）：

apt install zlib1g zlib1g-dev libxml2 libxml2-dev  # ubuntu
cd python; python3 -m pip install -r requirements.txt

构建安装（网络畅通环境下推荐使用）：

cd python
export FLAGTREE_BACKEND=backendxxx
python3 -m pip install . --no-build-isolation -v

构建技巧

自动下载依赖库的速度可能受限于网络环境，编译前可自行下载至缓存目录 ~/.flagtree（可通过环境变量 FLAGTREE_CACHE_DIR 修改），无需自行设置 LLVM_BUILD_DIR 等环境变量。
各后端完整构建命令如下：

iluvatar

# 推荐使用镜像 Ubuntu 20.04
mkdir -p ~/.flagtree/iluvatar; cd ~/.flagtree/iluvatar
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/iluvatar-llvm18-x86_64_v0.3.0.tar.gz
tar zxvf iluvatar-llvm18-x86_64_v0.3.0.tar.gz
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/iluvatarTritonPlugin-cpython3.10-glibc2.30-glibcxx3.4.28-cxxabi1.3.12-ubuntu-x86_64_v0.3.0.tar.gz
tar zxvf iluvatarTritonPlugin-cpython3.10-glibc2.30-glibcxx3.4.28-cxxabi1.3.12-ubuntu-x86_64_v0.3.0.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/python
export FLAGTREE_BACKEND=iluvatar
python3 -m pip install . --no-build-isolation -v

xpu (klx)

# 推荐使用镜像（22GB）https://su.bcebos.com/klx-sdk-release-public/xpytorch/docker/ubuntu2004_v030/ubuntu_2004_x86_64_v30.tar
# 联系 kunlunxin-support@baidu.com 可获取进一步支持
mkdir -p ~/.flagtree/xpu; cd ~/.flagtree/xpu
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/XTDK-llvm19-ubuntu2004_x86_64_v0.3.0.tar.gz
tar zxvf XTDK-llvm19-ubuntu2004_x86_64_v0.3.0.tar.gz
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/xre-Linux-x86_64_v0.3.0.tar.gz
tar zxvf xre-Linux-x86_64_v0.3.0.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/python
export FLAGTREE_BACKEND=xpu
python3 -m pip install . --no-build-isolation -v

mthreads

# 推荐使用镜像 flagtree/dockerfiles/Dockerfile-ubuntu22.04-python3.10-mthreads
mkdir -p ~/.flagtree/mthreads; cd ~/.flagtree/mthreads
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/mthreads-llvm19-glibc2.34-glibcxx3.4.30-x64_v0.1.0.tar.gz
tar zxvf mthreads-llvm19-glibc2.34-glibcxx3.4.30-x64_v0.1.0.tar.gz
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/mthreadsTritonPlugin-cpython3.10-glibc2.35-glibcxx3.4.30-cxxabi1.3.13-ubuntu-x86_64_v0.3.0.tar.gz
tar zxvf mthreadsTritonPlugin-cpython3.10-glibc2.35-glibcxx3.4.30-cxxabi1.3.13-ubuntu-x86_64_v0.3.0_v0.3.0.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/python
export FLAGTREE_BACKEND=mthreads
python3 -m pip install . --no-build-isolation -v

aipu (arm npu)

# 推荐使用镜像 Ubuntu 22.04
mkdir -p ~/.flagtree/aipu; cd ~/.flagtree/aipu
# 模拟环境中使用 x64 版本，在 ARM 开发板上使用 arm64 版本
wget https://oaitriton.blob.core.windows.net/public/llvm-builds/llvm-a66376b0-ubuntu-x64.tar.gz
tar zxvf llvm-a66376b0-ubuntu-x64.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/
git checkout -b triton_v3.3.x origin/triton_v3.3.x
export FLAGTREE_BACKEND=aipu
python3 -m pip install . --no-build-isolation -v

tsingmicro

# 推荐使用镜像 Ubuntu 20.04
mkdir -p ~/.flagtree/tsingmicro; cd ~/.flagtree/tsingmicro
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/tsingmicro-llvm21-glibc2.30-glibcxx3.4.28-python3.11-x64_v0.2.0.tar.gz
tar zxvf tsingmicro-llvm21-glibc2.30-glibcxx3.4.28-python3.11-x64_v0.2.0.tar.gz
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/tx8_depends_release_20250814_195126_v0.2.0.tar.gz
tar zxvf tx8_depends_release_20250814_195126_v0.2.0.tar.gz
export TX8_DEPS_ROOT=~/.flagtree/tsingmicro/tx8_deps
cd ${YOUR_CODE_DIR}/flagtree/
git checkout -b triton_v3.3.x origin/triton_v3.3.x
export FLAGTREE_BACKEND=tsingmicro
python3 -m pip install . --no-build-isolation -v

ascend

# 推荐使用镜像 flagtree/dockerfiles/Dockerfile-ubuntu22.04-python3.11-ascend
# 在 https://www.hiascend.com/developer/download/community/result?module=cann
# 注册账号后下载对应平台的 cann-toolkit、cann-kernels
# cann-toolkit
chmod +x Ascend-cann-toolkit_8.3.RC1.alpha001_linux-aarch64.run
./Ascend-cann-toolkit_8.3.RC1.alpha001_linux-aarch64.run --install
# cann-kernels for 910B (A2)
chmod +x Ascend-cann-kernels-910b_8.3.RC1.alpha001_linux-aarch64.run
./Ascend-cann-kernels-910b_8.3.RC1.alpha001_linux-aarch64.run --install
# cann-kernels for 910C (A3)
chmod +x Atlas-A3-cann-kernels_8.3.RC1.alpha001_linux-aarch64.run
./Atlas-A3-cann-kernels_8.3.RC1.alpha001_linux-aarch64.run --install
# 构建安装
mkdir -p ~/.flagtree/ascend; cd ~/.flagtree/ascend
wget https://oaitriton.blob.core.windows.net/public/llvm-builds/llvm-b5cc222d-ubuntu-arm64.tar.gz
tar zxvf llvm-b5cc222d-ubuntu-arm64.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/python
git checkout -b triton_v3.2.x origin/triton_v3.2.x
export FLAGTREE_BACKEND=ascend
python3 -m pip install . --no-build-isolation -v

hcu

# 推荐使用镜像 flagtree/dockerfiles/Dockerfile-ubuntu22.04-python3.10-hcu
mkdir -p ~/.flagtree/hcu; cd ~/.flagtree/hcu
wget baai-cp-web.ks3-cn-beijing.ksyuncs.com/trans/hcu-llvm20-df0864e-glibc2.35-glibcxx3.4.30-ubuntu-x86_64_v0.3.0.tar.gz
tar zxvf hcu-llvm20-df0864e-glibc2.35-glibcxx3.4.30-ubuntu-x86_64_v0.3.0.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/python
export FLAGTREE_BACKEND=hcu
python3 -m pip install . --no-build-isolation -v

nvidia 使用默认的构建命令，可以构建安装 nvidia、amd、triton_shared cpu 后端：

cd ${YOUR_LLVM_DOWNLOAD_DIR}
wget https://oaitriton.blob.core.windows.net/public/llvm-builds/llvm-10dc3a8e-ubuntu-x64.tar.gz
tar zxvf llvm-10dc3a8e-ubuntu-x64.tar.gz
cd ${YOUR_CODE_DIR}/flagtree/python
export LLVM_SYSPATH=${YOUR_LLVM_DOWNLOAD_DIR}/llvm-10dc3a8e-ubuntu-x64
export LLVM_INCLUDE_DIRS=$LLVM_SYSPATH/include
export LLVM_LIBRARY_DIR=$LLVM_SYSPATH/lib
unset FLAGTREE_BACKEND
python3 -m pip install . --no-build-isolation -v
# 如果接下来需要构建安装其他后端，应清空 LLVM 相关环境变量
unset LLVM_SYSPATH LLVM_INCLUDE_DIRS LLVM_LIBRARY_DIR

运行测试

安装完成后可以在设备支持的环境下运行测试：

# nvidia
cd python/test
python3 -m pytest -s
# other backends
cd third_party/backendxxx/python/test
python3 -m pytest -s
python3 test_xxx.py

关于贡献

欢迎参与 FlagTree 的开发并贡献代码，详情请参考CONTRIBUTING.md。

许可证

FlagTree 使用 MIT license。