From 2007f630b47d52738b3896ec7d6af90c50b129d2 Mon Sep 17 00:00:00 2001 From: Luo Tao Date: Thu, 31 May 2018 19:23:24 +0800 Subject: [PATCH] add build and install document of fluid inference library --- doc/fluid/howto/index_cn.rst | 2 +- doc/fluid/howto/index_en.rst | 1 - .../inference/build_and_install_lib_cn.rst | 96 +++++++++++++++++++ doc/fluid/howto/inference/index_cn.rst | 8 ++ ...id.md => inference_support_in_fluid_cn.md} | 59 +----------- 5 files changed, 106 insertions(+), 60 deletions(-) create mode 100644 doc/fluid/howto/inference/build_and_install_lib_cn.rst create mode 100644 doc/fluid/howto/inference/index_cn.rst rename doc/fluid/howto/inference/{inference_support_in_fluid.md => inference_support_in_fluid_cn.md} (90%) diff --git a/doc/fluid/howto/index_cn.rst b/doc/fluid/howto/index_cn.rst index b7c6201797..b57af64f44 100644 --- a/doc/fluid/howto/index_cn.rst +++ b/doc/fluid/howto/index_cn.rst @@ -4,5 +4,5 @@ .. toctree:: :maxdepth: 1 + inference/index_cn.rst optimization/index_cn.rst - inference/inference_support_in_fluid.md diff --git a/doc/fluid/howto/index_en.rst b/doc/fluid/howto/index_en.rst index f3ca41cdbf..fd21e167ce 100644 --- a/doc/fluid/howto/index_en.rst +++ b/doc/fluid/howto/index_en.rst @@ -5,4 +5,3 @@ HOW TO :maxdepth: 1 optimization/index_en.rst - inference/inference_support_in_fluid.md diff --git a/doc/fluid/howto/inference/build_and_install_lib_cn.rst b/doc/fluid/howto/inference/build_and_install_lib_cn.rst new file mode 100644 index 0000000000..c8d9992fcc --- /dev/null +++ b/doc/fluid/howto/inference/build_and_install_lib_cn.rst @@ -0,0 +1,96 @@ +安装与编译C++预测库 +=========================== + +直接下载安装 +------------- + +====================== ======================================== +版本说明 C++预测库 +====================== ======================================== +cpu_avx_mkl `fluid.tgz `_ +cpu_avx_openblas `fluid.tgz `_ +cpu_noavx_openblas `fluid.tgz `_ +cuda7.5_cudnn5_avx_mkl `fluid.tgz `_ +cuda8.0_cudnn5_avx_mkl `fluid.tgz `_ +cuda8.0_cudnn7_avx_mkl `fluid.tgz `_ +====================== ======================================== + +从源码编译 +---------- +用户也可以从 PaddlePaddle 核心代码编译C++预测库，只需在编译时配制下面这些编译选项： + +================= ========= +选项 值 +================= ========= +CMAKE_BUILD_TYPE Release +FLUID_INSTALL_DIR 安装路径 +WITH_FLUID_ONLY ON（推荐） +WITH_SWIG_PY OFF（推荐 +WITH_PYTHON OFF（推荐） +WITH_GPU ON/OFF +WITH_MKL ON/OFF +================= ========= + +建议按照推荐值设置，以避免链接不必要的库。其它可选编译选项按需进行设定。 + +下面的代码片段从github拉取最新代码，配制编译选项（需要将PADDLE_ROOT替换为PaddlePaddle预测库的安装路径）： + + .. code-block:: bash + + pip install paddlepaddle-gpu + PADDLE_ROOT=/path/of/capi + git clone https://github.com/PaddlePaddle/Paddle.git + cd Paddle + mkdir build + cd build + cmake -DFLUID_INSTALL_DIR=$PADDLE_ROOT \ + -DCMAKE_BUILD_TYPE=Release \ + -DWITH_FLUID_ONLY=ON \ + -DWITH_SWIG_PY=OFF \ + -DWITH_PYTHON=OFF \ + -DWITH_MKL=OFF \ + -DWITH_GPU=OFF \ + .. + make + make inference_lib_dist + +成功编译后，使用C++预测库所需的依赖（包括：（1）编译出的PaddlePaddle预测库和头文件；（2）第三方链接库和头文件；（3）版本信息与编译选项信息） +均会存放于PADDLE_ROOT目录中。目录结构如下： + + .. code-block:: text + + PaddleRoot/ + ├── CMakeCache.txt + ├── paddle + │   └── fluid + │   ├── framework + │   ├── inference + │   ├── memory + │   ├── platform + │   ├── pybind + │   └── string + ├── third_party + │   ├── boost + │   │   └── boost + │   ├── eigen3 + │   │   ├── Eigen + │   │   └── unsupported + │   └── install + │   ├── gflags + │   ├── glog + │   ├── mklml + │   ├── protobuf + │   ├── snappy + │   ├── snappystream + │   └── zlib + └── version.txt + +version.txt 中记录了该预测库的版本信息，包括Git Commit ID、使用OpenBlas或MKL数学库、CUDA/CUDNN版本号，如： + + .. code-block:: text + + GIT COMMIT ID: c95cd4742f02bb009e651a00b07b21c979637dc8 + WITH_MKL: ON + WITH_GPU: ON + CUDA version: 8.0 + CUDNN version: v5 diff --git a/doc/fluid/howto/inference/index_cn.rst b/doc/fluid/howto/inference/index_cn.rst new file mode 100644 index 0000000000..a903423548 --- /dev/null +++ b/doc/fluid/howto/inference/index_cn.rst @@ -0,0 +1,8 @@ +预测库 +------------ + +.. toctree:: + :maxdepth: 1 + + build_and_install_lib_cn.rst + inference_support_in_fluid_cn.md diff --git a/doc/fluid/howto/inference/inference_support_in_fluid.md b/doc/fluid/howto/inference/inference_support_in_fluid_cn.md similarity index 90% rename from doc/fluid/howto/inference/inference_support_in_fluid.md rename to doc/fluid/howto/inference/inference_support_in_fluid_cn.md index d272cd3e3b..309b17fccd 100644 --- a/doc/fluid/howto/inference/inference_support_in_fluid.md +++ b/doc/fluid/howto/inference/inference_support_in_fluid_cn.md @@ -1,9 +1,8 @@ -# Fluid Inference使用指南 +# 使用指南 ## 目录： - Python Inference API -- 编译Fluid Inference库 - Inference C++ API - Inference实例 - Inference计算优化 @@ -55,62 +54,6 @@ return [program, feed_target_names, fetch_targets] ``` - -## 编译Fluid Inference库 - - - **不需要额外的CMake选项** - - 1、 配置CMake命令，更多配置请参考[源码编译PaddlePaddle](http://www.paddlepaddle.org/docs/develop/documentation/zh/build_and_install/build_from_source_cn.html) - ```bash - $ git clone https://github.com/PaddlePaddle/Paddle.git - $ cd Paddle - $ mkdir build - $ cd build - $ cmake -DCMAKE_INSTALL_PREFIX=your/path/to/paddle_inference_lib \ - -DCMAKE_BUILD_TYPE=Release \ - -DWITH_PYTHON=ON \ - -DWITH_MKL=OFF \ - -DWITH_GPU=OFF \ - .. - ``` - - - 2、 编译PaddlePaddle - ```bash - $ make - ``` - - - 3、 部署。执行如下命令将PaddlePaddle Fluid Inference库部署到`your/path/to/paddle_inference_lib`目录。 - ```bash - $ make inference_lib_dist - ``` - -- 目录结构 - - ```bash - $ cd your/path/to/paddle_inference_lib - $ tree - . - |-- paddle - | `-- fluid - | |-- framework - | |-- inference - | | |-- io.h - | | `-- libpaddle_fluid.so - | |-- memory - | |-- platform - | `-- string - |-- third_party - | |-- eigen3 - | `-- install - | |-- gflags - | |-- glog - | `-- protobuf - `-- ... - ``` - - 假设`PADDLE_ROOT=your/path/to/paddle_inference_lib`。 - - - ## 链接Fluid Inference库 - 示例项目([链接](https://github.com/luotao1/fluid_inference_example.git)) -- GitLab