Compile on Windows from Source Code¶

Environment preparation¶

Windows 7/8/10 Pro/Enterprise(64bit)
GPU Version support CUDA 10.1/10.2/11.0/11.2, and only support single GPU
Python version 3.6+/3.7+/3.8+/3.9+(64bit)
pip version 20.2.2 or above (64bit)
Visual Studio 2015 Update3

Choose CPU/GPU¶

If your computer doesn’t have NVIDIA® GPU, please install CPU version of PaddlePaddle
If your computer has NVIDIA® GPU, and the following conditions are met，GPU version of PaddlePaddle is recommended.
- CUDA toolkit 9.0/10.0/10.1/10.2 with cuDNN 7 (cuDNN version>=7.6.5）
- CUDA toolkit 11.0 with cuDNN v8.0.4
- CUDA toolkit 11.2 with cuDNN v8.1.1
- GPU’s computing capability exceeds 3.5

Installation steps¶

There is one compilation methods in Windows system:

Direct native source code compilation(NCCL, distributed and other related functions are not supported temporarily)

Direct native source code compilation¶

Please strictly follow the following instructions step by step

Install the necessary tools i.e. cmake, git and python:

CMake requires version 3.10 and above, but there are official Bug versions of 3.12/3.13/3.14 when the GPU is compiled, we recommend that you use CMake3. 16 version, available on the official website [download] (https://cmake.org/download/), and add to the ring Environment variables.

Python requires version 3.6 and above, which can be downloaded from the official website.
- After installing python, please check whether the python version is the expected version by python-version, because you may have more than one python installed on your computer. You can handle conflicts of multiple pythons by changing the order of the environment variables.
numpy, protobuf, wheel are needed to be installed. Use the ‘pip’ command.
- To Install numpy package you can use command
```
pip install numpy
```
- To Install protobuf package you can use command
```
pip install protobuf
```
- To Install Wheel package you can use command
```
pip install wheel
```
Git can be downloaded on the official website and added to the environment variable.
Clone the PaddlePaddle source code to the Paddle subdirectories of the current directory and go to the Paddle subdirectories:
```
git clone https://github.com/PaddlePaddle/Paddle.git
```
```
cd Paddle
```
Switch to a more stable release branch for compilation:
```
git checkout [name of the branch]
```
For example：
```
git checkout release/2.1
```
Note: python3.6、python3.7 version started supporting from release/1.2, python3.8 version started supporting from release/1.8, python3.9 version started supporting from release/2.1
Create a directory called build and enter it:
```
mkdir build
```
```
cd build
```
Execute cmake:

For details on the compilation options, see the compilation options list.
- For users who need to compile the CPU version PaddlePaddle:
```
cmake .. -G "Visual Studio 14 2015 Win64" -DWITH_GPU=OFF -DWITH_TESTING=OFF -DCMAKE_BUILD_TYPE=Release
```
- For users who need to compile the GPU version PaddlePaddle:
```
cmake .. -G "Visual Studio 14 2015 Win64" -DWITH_GPU=ON -DWITH_TESTING=OFF -DCMAKE_BUILD_TYPE=Release
```
Python3 please add：

-DPY_VERSION=3 (or 3.6、3.7、3.8、3.9)

If your device information contains multiple Python or CUDA, you can also specify a specific version of Python or CUDA by setting the corresponding compile options:

-DPYTHON_EXECUTABLE: the installation path of python

-DCUDA_TOOLKIT_ROOT_DIR: the installation path of CUDA

For example: (for instance only, please set it according to your actual installation path)
```
cmake .. -G "Visual Studio 14 2015 Win64" -DCMAKE_BUILD_TYPE=Release -DWITH_GPU=ON -DWITH_TESTING=OFF -DPYTHON_EXECUTABLE=C:\\Python36\\python.exe -DCUDA_TOOLKIT_ROOT_DIR="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\v10.0"
```
Use Visual Studio 2015 to open paddle.sln file, select the platform x64, configure with Release, then begin to compile
After compilation successfully, go to the \paddle\build\python\dist directory and find the generated .whl package:
```
cd \paddle\build\python\dist
```
Install the generated .whl package:
```
pip install -U (whl package name)
```

Congratulations, you have completed the process of compiling PaddlePaddle successfully!

Verify installation¶

After the compilation and installation is completed, you can use python to enter the Python interface, input

import paddle

and then

           paddle.utils.run_check()

          

to verify that the installation was successful.

If PaddlePaddle is installed successfully! appears, it means the compilation and installation was successful.

How to uninstall¶

Please use the following command to uninstall PaddlePaddle:

CPU version of PaddlePaddle :
```
pip uninstall paddlepaddle
```
GPU version of PaddlePaddle :
```
pip uninstall paddlepaddle-gpu
```