The following instructions have been tested under Windows 10 Pro.
Download Visual Studio Community Edition from from the official website visualstudio.microsoft.com The current build instructions are tested with Visual Studio 2019.
Download Qt from from the official website qt.io
Make sure to choose the open source version of Qt and download the online installer.
Select Qt 5.15.0 (LTS) and open the sub tree. Only ‘MSVC 2019 64-bit’ component is needed.
Make a symbolic link to the directory corresponding to the current Qt version; e.g., if Qt 5.15.0 is installed in C:\Qt\5.15.0, then in Windows cmd (with administrative rights) execute:
$ mklink /D "C:\Qt\current" "C:\Qt\5.15.0"
Download the Windows x64 Installer from the official website cmake.org/download
Next up, download Python 3.9. For that, go to python.org/downloads/, look for Python 3.9.x and download the Windows installer (64-bit). Follow the setup and make sure to add Python to PATH.
Once Python is set up correctly, install numpy. Therefore open a PowerShell and type the command:
pip install numpy
Analogously install matplotlib with
pip install matplotlib
Install SWIG from the official website swig.org/download.html
Information can be found on [http://www.swig.org/Doc1.3/windows.html#Windows_installation]
Unpack the ZIP file and add its folder to PATH (so f.e. C:\swigwin-4.0.2)
Go to https://computing.mlz-garching.de/download/WinLibs
From there download:
Create the folders
C:\opt\x64\include
C:\opt\x64\lib
and paste the corresponding content from the ZIP files in those two folders.
On top of that, download the latest ZIP from the official website eigen.tuxfamily.org and copy the Eigen folder to C:\opt\x64\include
Additional: Ninja
The time to compile can be significantly improved, depending on the hardware, if Ninja is used as a build system. To download Ninja, go to ninja-build.org and download the Windows binary via GitHub. Add the binaries directory in your
PATH, f.e.C:\Program Files\ninja.
After those installations your PATH should contain
C:\Qt\5.15.0\msvc2019_64\bin
C:\Program Files\Python39\Scripts\
C:\Program Files\Python39\
C:\opt\x64\include
C:\opt\x64\lib
C:\msys64\usr\bin
C:\Program Files\CMake\bin
With all prerequisites completed, you should be able to build and test BornAgain in a PowerShell. In there, execute
$OPTLIBS = "C:/opt/x64"
$FFTW3_INCLUDE_DIR = "$OPTLIBS/include"
$FFTW3_LIB = "$OPTLIBS/lib/libfftw3-3.lib"
$QTDIR = "C:/Qt/current/msvc2019_64"
$QTCMake_DIR = "$QTDIR/lib/cmake"
$BUILD_DIR = "build"
mkdir -Force "$BUILD_DIR"
cd "$BUILD_DIR"
cmake -G "Visual Studio 16 2019" -A x64 -T host=x64 -DLIB_MAN=OFF -DQTDIR="$QTDIR" -DQt5_DIR="$QTCMake_DIR/Qt5" -DQt5Test_DIR="$QTCMake_DIR/Qt5Test" -DFFTW3_INCLUDE_DIR="$FFTW3_INCLUDE_DIR" -DFFTW3_LIBRARY="$FFTW3_LIB" -DCMAKE_INCLUDE_PATH="$OPTLIBS/include" -DCMAKE_LIBRARY_PATH="$OPTLIBS/lib" -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DCMAKE_C_COMPILER="cl.exe" -DCMAKE_CXX_COMPILER="cl.exe" -B. ..
cmake --build . --config Release
ctest -C Release --parallel 8 --output-on-failure
Additional: Build with Ninja (Experimental)
To use Ninja as the build system, the compilation cannot be performed in a PowerShell but in the Visual Studio Shell x64 Native Tools Command Prompt for VS 2019 which you can serach for in the start menu.
set OPTLIBS=C:/opt/x64 set FFTW3_INCLUDE_DIR=%OPTLIBS%/include set FFTW3_LIB=%OPTLIBS%/lib/libfftw3-3.lib set QTDIR=C:/Qt/current/msvc2019_64 set QTCMake_DIR=%QTDIR%/lib/cmake set BUILD_DIR=buildnj mkdir "%BUILD_DIR%" cd "%BUILD_DIR%" cmake --version cmake -G "Ninja" -DCMAKE_BUILD_TYPE=Release -DQTDIR="%QTDIR%" -DQt5_DIR="%QTCMake_DIR%/Qt5" -DQt5Test_DIR="%QTCMake_DIR%/Qt5Test" -DFFTW3_INCLUDE_DIR="%FFTW3_INCLUDE_DIR%" -DFFTW3_LIBRARY="%FFTW3_LIB%" -DCMAKE_INCLUDE_PATH="%OPTLIBS%/include" -DCMAKE_LIBRARY_PATH="%OPTLIBS%/lib" -DLIB_MAN=OFF -DCMAKE_C_COMPILER="cl.exe" -DCMAKE_CXX_COMPILER="cl.exe" -B. .. echo "## BUILD START:" %time% cmake --build . --config Release --parallel 8 echo "## BUILD END:" %time% ctest -C Release --parallel 8 --output-on-failure