CUDA 11.0+
Compilation has been tested with Microsoft Visual Studio 19 (under Windows 10).
To compile the code, you'll need CMake-GUI 3.21+ (https://cmake.org) on Windows.
Step 1. Compile dependencies: In CMake-GUI, browse source code to .../simplex/proj/_install; browse build to .../simplex/build/_install/cmake-build-win; configue with VS16 2019 and default compiler; generate and open project. Turn to Release mode and build, then close the project.
Step 2. Compile opengl viewer: Repeat Step 1 by replacing _install with opengl_viewer.
Step 3. Compile project: In CMake_GUI, browse source to .../complex/proj/PainlessSolver/proj/fluid_topo; browse build to .../complex/build/fluid_topo/cmake-build-win; configure, generate and open project. Turn to Release mode and build.
Step 4. Copy files under .../simplex/bin/win/opengl_viewer/Release to .../complex/bin/win/fluid_topo/Release.
In the command tool, go to the .../complex/bin/win/fluid_topo/Release folder, type:
./fluid_topo.exe -driver 0 -test 0 -s 256 -o output -frac 0.15 -lf 40
Here, -s for resolution, -o for file name to write result, -frac for volume fraction, and -lf for optimization iteration.
To visualize the result, type:
./opengl_viewer.exe -m fluid_dec -o output
which will open a opengl window. Type p for play, g for hide/display background grid, v for hide/display velocity, x for hide/display density distribution.