next

Edit

Build SINGA on Windows

The process of building SINGA from source on Microsoft Windows has four parts: install dependencies, build SINGA source, (optionally) install the python module and (optionally) run the unit tests.

Install Dependencies

You may create a folder for building the dependencies.

The dependencies are:

  • Compiler and IDE

  • CMake

    • Can be downloaded from http://cmake.org/
    • Make sure the path to cmake executable is in the system path, or use full path when calling cmake.

  • SWIG

    • Can be downloaded from http://swig.org/
    • Make sure the path to swig executable is in the system path, or use full path when calling swig. Use a recent version such as 3.0.12.

  • Protocol Buffers

    • Download a suitable version such as 2.6.1: https://github.com/google/protobuf/releases/tag/v2.6.1 .
    • Download both protobuf-2.6.1.zip and protoc-2.6.1-win32.zip .
    • Extract both of them in dependencies folder. Add the path to protoc executable to the system path, or use full path when calling it.
    • Open the Visual Studio solution which can be found in vsproject folder.
    • Change the build settings to Release and x64.
    • build libprotobuf project.

  • Openblas

    • Download a suitable source version such as 0.2.20 from http://www.openblas.net
    • Extract the source in the dependencies folder.
    • If you don't have Perl installed, download a perl environment such as Strawberry Perl (http://strawberryperl.com/)
    • Build the Visual Studio solution by running this command in the source folder:
    cmake -G "Visual Studio 15 2017 Win64"
    • Open the Visual Studio solution and change the build settings to Release and x64.
    • Build libopenblas project

  • Google glog

    • Download a suitable version such as 0.3.5 from https://github.com/google/glog/releases
    • Extract the source in the dependencies folder.
    • Open the Visual Studio solution.
    • Change the build settings to Release and x64.
    • Build libglog project

Build SINGA source

  • Download SINGA source code

  • Compile the protobuf files:

    • Go to src/proto folder
    mkdir python_out
protoc.exe *.proto --python_out python_out

  • Generate swig interfaces for C++ and Python: Go to src/api

    swig -python -c++ singa.i

  • generate Visual Studio solution for SINGA: Go to SINGA source code root folder

    mkdir build
cd build

  • Call cmake and add the paths in your system similar to the following example:

    cmake -G "Visual Studio 15 2017 Win64" ^
  -DGLOG_INCLUDE_DIR="D:/WinSinga/dependencies/glog-0.3.5/src/windows" ^
  -DGLOG_LIBRARIES="D:/WinSinga/dependencies/glog-0.3.5/x64/Release" ^
  -DCBLAS_INCLUDE_DIR="D:/WinSinga/dependencies/openblas-0.2.20/lapack-netlib/CBLAS/include" ^
  -DCBLAS_LIBRARIES="D:/WinSinga/dependencies/openblas-0.2.20/lib/RELEASE" ^
  -DProtobuf_INCLUDE_DIR="D:/WinSinga/dependencies/protobuf-2.6.1/src" ^
  -DProtobuf_LIBRARIES="D:/WinSinga/dependencies/protobuf-2.6.1/vsprojects/x64/Release" ^
  -DProtobuf_PROTOC_EXECUTABLE="D:/WinSinga/dependencies/protoc-2.6.1-win32/protoc.exe" ^
  ..

  • Open the generated solution in Visual Studio

  • Change the build settings to Release and x64

  • Add the singa_wrap.cxx file from src/api to the singa_objects project

  • In the singa_objects project, open Additional Include Directories.

  • Add Python include path

  • Add numpy include path

  • Add protobuf include path

  • In the preprocessor definitions of the singa_objects project, add USE_GLOG

  • Build singa_objects project

  • In singa project:

    • add singa_wrap.obj to Object Libraries
    • change target name to _singa_wrap
    • change target extension to .pyd
    • change configuration type to Dynamic Library (.dll)
    • go to Additional Library Directories and add the path to python, openblas, protobuf and glog libraries
    • go to Additional Dependencies and add libopenblas.lib, libglog.lib and libprotobuf.lib

  • build singa project

Install Python module

  • Change _singa_wrap.so to _singa_wrap.pyd in build/python/setup.py

  • Copy the files in src/proto/python_out to build/python/singa/proto

  • Optionally create and activate a virtual environment:

    mkdir SingaEnv
virtualenv SingaEnv
SingaEnv\Scripts\activate

  • go to build/python folder and run:

    python setup.py install

  • Make _singa_wrap.pyd, libglog.dll and libopenblas.dll available by adding them to the path or by copying them to singa package folder in the python site-packages

  • Verify that SINGA is installed by running:

    python -c "from singa import tensor"

A video tutorial for the build process can be found here:

youtube video

Run Unit Tests

  • In the test folder, generate the Visual Studio solution:

    cmake -G "Visual Studio 15 2017 Win64"

  • Open the generated solution in Visual Studio.

  • Change the build settings to Release and x64.

  • Build glog project.

  • In test_singa project:

    • Add USE_GLOG to the Preprocessor Definitions.
    • In Additional Include Directories, add path of GLOG_INCLUDE_DIR, CBLAS_INCLUDE_DIR and Protobuf_INCLUDE_DIR which were used in step 2 above. Add also build and build/include folders.
    • Go to Additional Library Directories and add the path to openblas, protobuf and glog libraries. Add also build/src/singa_objects.dir/Release.
    • Go to Additional Dependencies and add libopenblas.lib, libglog.lib and libprotobuf.lib. Fix the names of the two libraries: gtest.lib and singa_objects.lib.

  • Build test_singa project.

  • Make libglog.dll and libopenblas.dll available by adding them to the path or by copying them to test/release folder

  • The unit tests can be executed

    • From the command line:
    test_singa.exe
    • From Visual Studio:
      • right click on the test_singa project and choose 'Set as StartUp Project'.
      • from the Debug menu, choose 'Start Without Debugging'

A video tutorial for running the unit tests can be found here:

youtube video

Build GPU support with CUDA

In this section, we will extend the previous steps to enable GPU.

Install Dependencies

In addition to the dependencies in section 1 above, we will need the following:

Build SINGA source

  • Call cmake and add the paths in your system similar to the following example: 
    cmake -G "Visual Studio 15 2017 Win64" ^
  -DGLOG_INCLUDE_DIR="D:/WinSinga/dependencies/glog-0.3.5/src/windows" ^
  -DGLOG_LIBRARIES="D:/WinSinga/dependencies/glog-0.3.5/x64/Release" ^
  -DCBLAS_INCLUDE_DIR="D:/WinSinga/dependencies/openblas-0.2.20/lapack-netlib/CBLAS/include" ^
  -DCBLAS_LIBRARIES="D:/WinSinga/dependencies/openblas-0.2.20/lib/RELEASE" ^
  -DProtobuf_INCLUDE_DIR="D:/WinSinga/dependencies/protobuf-2.6.1/src" ^
  -DProtobuf_LIBRARIES="D:\WinSinga/dependencies/protobuf-2.6.1/vsprojects/x64/Release" ^
  -DProtobuf_PROTOC_EXECUTABLE="D:/WinSinga/dependencies/protoc-2.6.1-win32/protoc.exe" ^
  -DCUDNN_INCLUDE_DIR=D:\WinSinga\dependencies\cudnn-9.1-windows10-x64-v7.1\cuda\include ^
  -DCUDNN_LIBRARIES=D:\WinSinga\dependencies\cudnn-9.1-windows10-x64-v7.1\cuda\lib\x64 ^
  -DSWIG_DIR=D:\WinSinga\dependencies\swigwin-3.0.12 ^
  -DSWIG_EXECUTABLE=D:\WinSinga\dependencies\swigwin-3.0.12\swig.exe ^
  -DUSE_CUDA=YES ^
  -DCUDNN_VERSION=7 ^
  ..

  • Generate swig interfaces for C++ and Python: Go to src/api

    swig -python -c++ singa.i

  • Open the generated solution in Visual Studio

  • Change the build settings to Release and x64

Building singa_objects

  • Add the singa_wrap.cxx file from src/api to the singa_objects project
  • In the singa_objects project, open Additional Include Directories.
  • Add Python include path
  • Add numpy include path
  • Add protobuf include path
  • Add include path for CUDA, cuDNN and cnmem
  • In the preprocessor definitions of the singa_objects project, add USE_GLOG, USE_CUDA and USE_CUDNN. Remove DISABLE_WARNINGS.
  • Build singa_objects project

Building singa-kernel

  • Create a new Visual Studio project of type "CUDA 9.1 Runtime". Give it a name such as singa-kernel.

  • The project comes with an initial file called kernel.cu. Remove this file from the project.

  • Add this file: src/core/tensor/math_kernel.cu

  • In the project settings:

    • Set Platform Toolset to "Visual Studio 2015 (v140)"
    • Set Configuration Type to " Static Library (.lib)"
    • In the Include Directories, add build/include.

  • Build singa-kernel project

Building singa

  • In singa project:

    • add singa_wrap.obj to Object Libraries
    • change target name to _singa_wrap
    • change target extension to .pyd
    • change configuration type to Dynamic Library (.dll)
    • go to Additional Library Directories and add the path to python, openblas, protobuf and glog libraries
    • Add also the library path to singa-kernel, cnmem, cuda and cudnn.
    • go to Additional Dependencies and add libopenblas.lib, libglog.lib and libprotobuf.lib.
    • Add also: singa-kernel.lib, cnmem.lib, cudnn.lib, cuda.lib , cublas.lib, curand.lib and cudart.lib.

  • build singa project

Install Python module

  • Change _singa_wrap.so to _singa_wrap.pyd in build/python/setup.py

  • Copy the files in src/proto/python_out to build/python/singa/proto

  • Optionally create and activate a virtual environment:

    mkdir SingaEnv
virtualenv SingaEnv
SingaEnv\Scripts\activate

  • go to build/python folder and run:

    python setup.py install

  • Make _singa_wrap.pyd, libglog.dll, libopenblas.dll, cnmem.dll, CUDA Runtime (e.g. cudart64_91.dll) and cuDNN (e.g. cudnn64_7.dll) available by adding them to the path or by copying them to singa package folder in the python site-packages

  • Verify that SINGA is installed by running:

    python -c "from singa import device; dev = device.create_cuda_gpu()"

A video tutorial for this part can be found here:

youtube video

Run Unit Tests

  • In the test folder, generate the Visual Studio solution:

    cmake -G "Visual Studio 15 2017 Win64"

  • Open the generated solution in Visual Studio, or add the project to the singa solution that was created in step 5.2

  • Change the build settings to Release and x64.

  • Build glog project.

  • In test_singa project:

    • Add USE_GLOG; USE_CUDA; USE_CUDNN to the Preprocessor Definitions.
    • In Additional Include Directories, add path of GLOG_INCLUDE_DIR, CBLAS_INCLUDE_DIR and Protobuf_INCLUDE_DIR which were used in step 5.2 above. Add also build, build/include, CUDA and cuDNN include folders.
    • Go to Additional Library Directories and add the path to openblas, protobuf and glog libraries. Add also build/src/singa_objects.dir/Release, singa-kernel, cnmem, CUDA and cuDNN library paths.
    • Go to Additional Dependencies and add libopenblas.lib; libglog.lib; libprotobuf.lib; cnmem.lib; cudnn.lib; cuda.lib; cublas.lib; curand.lib; cudart.lib; singa-kernel.lib. Fix the names of the two libraries: gtest.lib and singa_objects.lib.

  • Build test_singa project.

  • Make libglog.dll, libopenblas.dll, cnmem.dll, cudart64_91.dll and cudnn64_7.dll available by adding them to the path or by copying them to test/release folder

  • The unit tests can be executed

    • From the command line:

      test_singa.exe

    • From Visual Studio:

      • right click on the test_singa project and choose 'Set as StartUp Project'.
      • from the Debug menu, choose 'Start Without Debugging'

A video tutorial for running the unit tests can be found here:

youtube video

Last updated on 10/4/2020