[DOC] Update build command in readme (#33)

Author: zhzhcookie

README.md

@@ -2,7 +2,7 @@
 
 ## FlagTree
 
-FlagTree is an open source, unified compiler for multiple AI chips project dedicated to developing a diverse ecosystem of AI chip compilers and related tooling platforms, thereby fostering and strengthening the upstream and downstream Triton ecosystem. Currently in its initial phase, the project aims to maintain compatibility with existing adaptation solutions while unifying the codebase to rapidly implement single-version multi-backend support.
+FlagTree is an open source, unified compiler for multiple AI chips project dedicated to developing a diverse ecosystem of AI chip compilers and related tooling platforms, thereby fostering and strengthening the upstream and downstream Triton ecosystem. Currently in its initial phase, the project aims to maintain compatibility with existing adaptation solutions while unifying the codebase to rapidly implement single-repository multi-backend support. For upstream model users, it provides unified compilation capabilities across multiple backends; for downstream chip manufacturers, it offers examples of Triton ecosystem integration.
 
 ## Install from source
 Installation dependencies (ensure you use the correct python3.x version):
@@ -28,7 +28,9 @@ Complete build commands for each backend:
 # Recommended: Use Ubuntu 20.04
 mkdir -p ~/.flagtree/iluvatar; cd ~/.flagtree/iluvatar
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/iluvatar-llvm18-x86_64.tar.gz
+tar zxvf iluvatar-llvm18-x86_64.tar.gz
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/iluvatarTritonPlugin-cpython3.10-glibc2.30-glibcxx3.4.28-cxxabi1.3.12-ubuntu-x86_64.tar.gz
+tar zxvf iluvatarTritonPlugin-cpython3.10-glibc2.30-glibcxx3.4.28-cxxabi1.3.12-ubuntu-x86_64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/python
 export FLAGTREE_BACKEND=iluvatar
 python3 -m pip install . --no-build-isolation -v
@@ -39,7 +41,9 @@ python3 -m pip install . --no-build-isolation -v
 # Contact kunlunxin-support@baidu.com for support
 mkdir -p ~/.flagtree/xpu; cd ~/.flagtree/xpu
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/XTDK-llvm19-ubuntu2004_x86_64.tar.gz
+tar zxvf XTDK-llvm19-ubuntu2004_x86_64.tar.gz
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/xre-Linux-x86_64.tar.gz
+tar zxvf xre-Linux-x86_64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/python
 export FLAGTREE_BACKEND=xpu
 python3 -m pip install . --no-build-isolation -v
@@ -49,6 +53,7 @@ python3 -m pip install . --no-build-isolation -v
 # Recommended: Use the Dockerfile flagtree/dockerfiles/Dockerfile-ubuntu22.04-python3.10-mthreads
 mkdir -p ~/.flagtree/mthreads; cd ~/.flagtree/mthreads
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/mthreads-llvm19-glibc2.34-glibcxx3.4.30-x64.tar.gz
+tar zxvf mthreads-llvm19-glibc2.34-glibcxx3.4.30-x64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/python
 export FLAGTREE_BACKEND=mthreads
 python3 -m pip install . --no-build-isolation -v
@@ -58,6 +63,7 @@ python3 -m pip install . --no-build-isolation -v
 # Recommended: Use Ubuntu 20.04
 mkdir -p ~/.flagtree/aipu; cd ~/.flagtree/aipu
 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
@@ -68,6 +74,7 @@ python3 -m pip install . --no-build-isolation -v
 # Recommended: Use Ubuntu 20.04
 mkdir -p ~/.flagtree/tsingmicro; cd ~/.flagtree/tsingmicro
 wget https://github.com/FlagTree/flagtree/releases/download/v0.2.0-build-deps/tsingmicro-llvm21-glibc2.30-glibcxx3.4.28-x64.tar.gz
+tar zxvf tsingmicro-llvm21-glibc2.30-glibcxx3.4.28-x64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/
 git checkout -b triton_v3.3.x origin/triton_v3.3.x
 export FLAGTREE_BACKEND=tsingmicro
@@ -86,28 +93,28 @@ chmod +x Atlas-A3-cann-kernels_8.1.RC1_linux-aarch64.run
 # build
 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
 ```
 
-To build with default backends (nvidia, amd, triton_shared):
+To build with default backends nvidia, amd, triton_shared (cpu):
 ```shell
 # manually download LLVM
 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
+tar zxvf llvm-10dc3a8e-ubuntu-x64.tar.gz
 # build
 cd ${YOUR_CODE_DIR}/flagtree/python
-export LLVM_BUILD_DIR=${YOUR_LLVM_DOWNLOAD_DIR}/llvm-10dc3a8e-ubuntu-x64
-export LLVM_INCLUDE_DIRS=$LLVM_BUILD_DIR/include
-export LLVM_LIBRARY_DIR=$LLVM_BUILD_DIR/lib
-export LLVM_SYSPATH=$LLVM_BUILD_DIR
+export LLVM_SYSPATH=${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
 # If you need to build other backends afterward, you should clear LLVM-related environment variables
-unset LLVM_BUILD_DIR LLVM_INCLUDE_DIRS LLVM_LIBRARY_DIR LLVM_SYSPATH
+unset LLVM_SYSPATH LLVM_INCLUDE_DIRS LLVM_LIBRARY_DIR
 ```
 
 ## Running tests

README_cn.md

@@ -2,7 +2,7 @@
 
 ## 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，快速实现单版本多后端支持。
+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）：
@@ -20,15 +20,17 @@ 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。
-各后端完整编译命令如下：
+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。 <br>
+各后端完整编译命令如下： <br>
 
 [iluvatar](/third_party/iluvatar/)
 ```shell
 # 推荐使用镜像 Ubuntu 20.04
 mkdir -p ~/.flagtree/iluvatar; cd ~/.flagtree/iluvatar
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/iluvatar-llvm18-x86_64.tar.gz
+tar zxvf iluvatar-llvm18-x86_64.tar.gz
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/iluvatarTritonPlugin-cpython3.10-glibc2.30-glibcxx3.4.28-cxxabi1.3.12-ubuntu-x86_64.tar.gz
+tar zxvf iluvatarTritonPlugin-cpython3.10-glibc2.30-glibcxx3.4.28-cxxabi1.3.12-ubuntu-x86_64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/python
 export FLAGTREE_BACKEND=iluvatar
 python3 -m pip install . --no-build-isolation -v
@@ -39,7 +41,9 @@ python3 -m pip install . --no-build-isolation -v
 # 联系 kunlunxin-support@baidu.com 可获取进一步支持
 mkdir -p ~/.flagtree/xpu; cd ~/.flagtree/xpu
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/XTDK-llvm19-ubuntu2004_x86_64.tar.gz
+tar zxvf XTDK-llvm19-ubuntu2004_x86_64.tar.gz
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/xre-Linux-x86_64.tar.gz
+tar zxvf xre-Linux-x86_64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/python
 export FLAGTREE_BACKEND=xpu
 python3 -m pip install . --no-build-isolation -v
@@ -49,6 +53,7 @@ python3 -m pip install . --no-build-isolation -v
 # 推荐使用镜像 flagtree/dockerfiles/Dockerfile-ubuntu22.04-python3.10-mthreads
 mkdir -p ~/.flagtree/mthreads; cd ~/.flagtree/mthreads
 wget https://github.com/FlagTree/flagtree/releases/download/v0.1.0-build-deps/mthreads-llvm19-glibc2.34-glibcxx3.4.30-x64.tar.gz
+tar zxvf mthreads-llvm19-glibc2.34-glibcxx3.4.30-x64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/python
 export FLAGTREE_BACKEND=mthreads
 python3 -m pip install . --no-build-isolation -v
@@ -58,6 +63,7 @@ python3 -m pip install . --no-build-isolation -v
 # 推荐使用镜像 Ubuntu 20.04
 mkdir -p ~/.flagtree/aipu; cd ~/.flagtree/aipu
 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
@@ -68,6 +74,7 @@ python3 -m pip install . --no-build-isolation -v
 # 推荐使用镜像 Ubuntu 20.04
 mkdir -p ~/.flagtree/tsingmicro; cd ~/.flagtree/tsingmicro
 wget https://github.com/FlagTree/flagtree/releases/download/v0.2.0-build-deps/tsingmicro-llvm21-glibc2.30-glibcxx3.4.28-x64.tar.gz
+tar zxvf tsingmicro-llvm21-glibc2.30-glibcxx3.4.28-x64.tar.gz
 cd ${YOUR_CODE_DIR}/flagtree/
 git checkout -b triton_v3.3.x origin/triton_v3.3.x
 export FLAGTREE_BACKEND=tsingmicro
@@ -77,35 +84,37 @@ python3 -m pip install . --no-build-isolation -v
 ```shell
 # 推荐使用镜像 flagtree/dockerfiles/Dockerfile-ubuntu20.04-python3.9-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，这里以 AArch64 架构的 A3 处理器为例展示如何安装
+# 注册账号后下载对应平台的 cann-toolkit、cann-kernels
+# 这里以 AArch64 架构的 A3 处理器为例展示如何安装
 chmod +x Ascend-cann-toolkit_8.2.RC1.alpha002_linux-aarch64.run
 ./Ascend-cann-toolkit_8.2.RC1.alpha002_linux-aarch64.run --install
 chmod +x Atlas-A3-cann-kernels_8.1.RC1_linux-aarch64.run
 ./Atlas-A3-cann-kernels_8.1.RC1_linux-aarch64.run --install
 # 编译安装
 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
 ```
 
-使用默认的编译命令，可以编译安装 nvidia、amd、triton_shared 后端：
+使用默认的编译命令，可以编译安装 nvidia、amd、triton_shared (cpu) 后端：
 ```shell
 # 自行下载 llvm
 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
+tar zxvf llvm-10dc3a8e-ubuntu-x64.tar.gz
 # 编译安装
 cd ${YOUR_CODE_DIR}/flagtree/python
-export LLVM_BUILD_DIR=${YOUR_LLVM_DOWNLOAD_DIR}/llvm-10dc3a8e-ubuntu-x64
-export LLVM_INCLUDE_DIRS=$LLVM_BUILD_DIR/include
-export LLVM_LIBRARY_DIR=$LLVM_BUILD_DIR/lib
-export LLVM_SYSPATH=$LLVM_BUILD_DIR
+export LLVM_SYSPATH=${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
 # 如果接下来需要编译安装其他后端，应清空 LLVM 相关环境变量
-unset LLVM_BUILD_DIR LLVM_INCLUDE_DIRS LLVM_LIBRARY_DIR LLVM_SYSPATH
+unset LLVM_SYSPATH LLVM_INCLUDE_DIRS LLVM_LIBRARY_DIR
 ```
 
 ## 运行测试

python/setup_tools/setup_helper.py

@@ -431,7 +431,7 @@ def check_env(env_val):
 
 # ascend
 cache.store(
-    file="ascend-llvm-b5cc222d-ubuntu-arm64",
+    file="llvm-b5cc222d-ubuntu-arm64",
     condition=("ascend" == flagtree_backend),
     url="https://oaitriton.blob.core.windows.net/public/llvm-builds/llvm-b5cc222d-ubuntu-arm64.tar.gz",
     pre_hock=lambda: check_env('LLVM_SYSPATH'),
@@ -440,7 +440,7 @@ def check_env(env_val):
 
 # aipu
 cache.store(
-    file="aipu-llvm-a66376b0-ubuntu-x64",
+    file="llvm-a66376b0-ubuntu-x64",
     condition=("aipu" == flagtree_backend),
     url="https://oaitriton.blob.core.windows.net/public/llvm-builds/llvm-a66376b0-ubuntu-x64.tar.gz",
     pre_hock=lambda: check_env('LLVM_SYSPATH'),

python/setup_tools/utils/__init__.py

@@ -17,10 +17,7 @@ class FlagTreeBackend:
                     tag="380b87122c88af131530903a702d5318ec59bb33"),
     FlagTreeBackend(name="cambricon", url="https://github.com/Cambricon/triton-linalg.git",
                     tag="00f51c2e48a943922f86f03d58e29f514def646d"),
-    FlagTreeBackend(
-        name="ascend",
-        url="https://gitee.com/ascend/triton-ascend.git",
-    ),
+    FlagTreeBackend(name="ascend", url="https://gitee.com/ascend/triton-ascend.git"),
 )