Visual Computing Library
•All Classes Functions Variables Typedefs Enumerations Friends Modules Pages Concepts
Installation

The library can be used both as a C++ library and as a Python library.

C++

In order to work easily with VCLib, you can use the CMake configuration, but you can also use the core module of the library as a header-only library.

CMake

The minimum required version is 3.24.

CMake will take care to find and download all the required and optional dependencies for each module of the library that you want to build and use. By default, the library is built with the Core module only, but you can enable the other modules by setting the VCLIB_BUILD_MODULE_<module> CMake variable to ON, before adding the VCLib directory to your project or in your CMake configuration.

There are two very easy ways to include VCLib in your project.

Using <em>add_subdirectory</em>

To include VCLib in your CMake project, first clone (recursively) the VCLib repository into your project directory:

cd /path/to/your/project
git clone --recursive https://github.com/cnr-isti-vclab/vclib

(alternatively, you can use git submodules).

Then, in your CMakeLists.txt:

# If you want to use other modules, set the corresponding variable to ON, e.g:
# set(VCLIB_BUILD_MODULE_RENDER ON)
add_subdirectory(vclib)
[...]
add_executable(my_target main.cpp)
target_link_libraries(my_target PRIVATE vclib::vclib)

Where my_target is the target of your project.

Using <em>FetchContent</em>

If you don't want to clone the repository, you can use FetchContent to download the library at configuration time. In your CMakeLists.txt:

include(FetchContent)

WORK IN PROGRESS

Usage with <em>INCLUDE_PATH</em>

You can also use the VCLib Core module just as a header-only library.

In order to work with VCLib, you must ensure that Eigen is accessible in your INCLUDE_PATH. MapBox Earcut, STB, and poolSTL are already included automatically by VCLib, so you don't need to include them manually. Eigen is also provided within the external directory of VCLib.

To set your INCLUDE_PATH correctly, follow these steps:

INCLUDE_PATH += path/to/vclib/external/eigen-3.4.0/ # or your favorite Eigen version
INCLUDE_PATH += path/to/vclib/include/

and you are ready to go.

Supported Compilers

VCLib requires recent compiler versions in order to build the newly features of the C++20 language, that have been recently integrated in the major compilers:

  • GCC 13.3 (default in Ubuntu 24.04)
  • CLang 16 (XCode 15.0 on MacOS)
  • MSVC 19.36 (Visual Studio 2022)

C++17 Parallel STL

The lack of support for C++17 parallel STL in some compilers is still a problem.

To allow the use of C++17 parallel STL, VCLib uses poolSTL as a fallback. poolSTL will use the parallel STL if it is available, and otherwise, it implements the parallel algorithms using a thread pool.

In the case of libstdc++, the parallel STL is available only if the Threading Building Blocks library is installed. If you are using libstdc++, we suggest to install TBB in order to bypass poolSTL and use the parallel STL.

On Ubuntu, you can install it by running the following command in the terminal:

sudo apt install libtbb-dev

Dependencies

VCLib has several dependencies that depend on the module you want to use. Along with TBB or poolSTL for parallelism, the Core module depends on the following libraries:

All these libraries are bundled with VCLib, so you don't need to install them separately.

All the other dependencies can be required or optional, and they are needed by the other modules of the library. The CMake configuration, when possible, will take care to download them for you using FetchContent. You can always enable or disable the usage of a dependency by setting the corresponding CMake variable called VCLIB_ALLOW_<dependency> to ON or OFF.

Python

WORK IN PROGRESS