Installation

Note

We require C++20, CMake v3.25.2+, and a recent C++ and CUDA compiler.

CMake Project

Integration with CMake is done in the usual ways:

1. Fetch the library.

   CPM

   If you use CPM.cmake, just add the following.

   CMakeLists.txt

   CPMAddPackage("gh:neilkichler/cuinterval@v0.2.1")

   FetchContent

   FetchContent is built-in to recent CMake. Ensure once at the top of your CMakelists.txt you have at least version 3.14 as a requirement.

   CMakelists.txt

   cmake_minimum_required(VERSION 3.14)

   And you need to include it somewhere.

   CMakeLists.txt

      include(FetchContent)

   Now all dependencies can be added using FetchContent_Declare and FetchContent_MakeAvailable. Here shown for cuinterval.

   CMakelists.txt

   FetchContent_Declare(
     cuinterval
     GIT_REPOSITORY git@github.com:neilkichler/cuinterval.git
     GIT_TAG main
   )
   FetchContent_MakeAvailable(cuinterval)

   System-wide

   A system-wide install is currently only possible by building from source.

   git clone https://github.com/neilkichler/cuinterval.git
   cd cuinterval
   cmake --preset release
   cmake --build build
   cmake --install build

   Note

   By default we make use of Ninja for building.

   You also need to add find_package(cuinterval) before linking to the library.

   CMakelists.txt

   find_package(cuinterval)

   Git submodule

   To manage the dependency as Git submodule, you can add it as follows.

     git submodule add https://github.com/neilkichler/cuinterval.git

   Note

   If you want to change things, a fork of the cuinterval library is probably your best option. Then, the submodule should link to your fork.

   Now you just need to add the subdirectory.

   CMakelists.txt

    add_subdirectory(path/to/cuinterval)

2. Add the library as a link dependency to your project.

   CMakelists.txt

   target_link_libraries(${PROJECT_NAME} PUBLIC cuinterval)

3. Take a look at the example to get started.

Header-only

You can directly include the header folder into your project.

The latest header-only version is available in the header-only branch: https://github.com/neilkichler/cuinterval/archive/header-only.zip

Note

Previous versions can be found in GitHub Releases.

Single Header

Download the latest amalgamated single header version of the library.

curl -o cuinterval.cuh https://raw.githubusercontent.com/neilkichler/cuinterval/refs/heads/single-header/cuinterval.cuh

Note

Previous versions can be found in GitHub Releases.