Compilation

LISFLOOD-FP can be compiled on Windows or Linux using CMake version 3.13 or above.

Compiling on Windows with the MSVC compiler

Launch Visual Studio 2019 and open lisflood-fp as a local folder. Choose the msvc-x64-Debug or msvc-x64-Release configuration, then Rebuild All.

Compiling on Windows with the Intel compiler

Close Visual Studio, then run launch_vs2019_intel64.bat, which assumes that Visual Studio 2019 Community Edition is installed (the .bat script sets the necessary environment variables for Visual Studio to locate the Intel compiler). Choose the intel-x64-Debug or intel-x64-Release configuration, then Rebuild All. Once launched from the .bat file, Visual Studio can compile with either MSVC or Intel compiler by choosing the appropriate configuration.

Running or debugging on Windows

To run lisflood.exe from Visual Studio, switch the Solution Explorer to CMake Targets View, the Add Debug Configuration to lisflood (executable) (see the Visual Studio documentation for more details).

Compiling on Linux

Ensure a recent version of CMake is installed. Also ensure libnuma-dev is installed and, for NetCDF output and dynamic rainfall support, ensure libnetcdf-dev is also installed. On Ubuntu:

sudo snap install cmake --classic
sudo apt install libnuma-dev libnetcdf-dev

Then open a terminal at in the root lisflood-fp directory:

cmake -S . -B build
cmake --build build

The lisflood executable is written to the build directory. If libnuma is installed in a non-standard location, use

cmake -S . -B build -DNUMA_ROOT=<path>
cmake --build build

Customising the build

The default build configuration is given in config.default.cmake. To customise the build, copy and modify this file. Then, in Windows, edit the appropriate configuration in CMakeSettings.json:

"configurations": [
  {
    "name": "msvc-x64-Debug",
    …
    "cmakeCommandArgs": "-D_CONFIG=<filename>",
    …
  }
  …
]

Or, in Linux:

cmake -S . -B build -D_CONFIG=<filename>
cmake --build build

NVIDIA CUDA support

CMake automatically compiles LISFLOOD-FP FV1 and DG2 CUDA solvers if the NVIDIA CUDA Toolkit is installed. To customise the CUDA compute capabilities, see config.default.cmake.