CMake C++/CUDA Multi-Platform Template for Visual Studio Code

This CMake project serves as a template for C++ and CUDA development. It includes the following directories and files:

• my_header_lib: C++ header-only library
• my_static_lib: C++ static library
• my_shared_lib: C++ shared library
• main: Executable
• test: Tests using Google Test
• cuda/cuda_main: (Optional) CUDA executable
• cuda/my_cuda_lib: (Optional) CUDA static library

Preconfigured linter and formatter settings based on the Google C++ Style Guide are included.

• .clang-format: ClangFormat 20
• .clang-tidy: Clang-Tidy 20
• .clang-tidy-lower: Lower-case version of the above. Please rename it if needed.

vcpkg is included as a submodule:

• vcpkg.json: List of dependencies

Setting Up the C++ Environment

Linux

Using apt

For Ubuntu 24.04, please install the following:

$ sudo apt install build-essential gdb g++-14 clang-20 clangd-20 cmake ninja-build pkg-config

Then, add the clangd path to the .vscode/settings.json file:

{
    "clangd.path": "/usr/bin/clangd-20"
}

Using Homebrew

If you prefer the latest compiler versions:

$ /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
$ brew install gcc llvm cmake ninja gdb

macOS

Using Apple Clang

$ xcode-select –install

Using Homebrew

$ /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
$ brew install gcc llvm cmake ninja gdb

Windows

Please install Visual Studio (or Build Tools for Visual Studio for command-line use) with the C++ development workload.

Other Dependencies

To generate compile_commands.json with header files, install compdb:

$ sudo apt install pipx
$ pipx install compdb

Note

Note that compdb currently does not support Windows.

Getting Started

Opening the Project Template

1. Clone this repository with --recursive to include submodules.

   $ git clone --recursive https://github.com/yosh-matsuda/vscode-cpp-cuda-cmake-template.git

2. Open the project in VSCode.

   $ code vscode-cpp-cuda-cmake-template

3. Please install the following VSCode extensions:

• clangd
• C/C++ Extension Pack

Building and Running

1. Select the compiler and build type in CMake Tools.

2. Press F7 to configure and build the project.

3. Press F1 and run clangd: Restart language server to load compile_commands.json and restart clangd.

4. Press F5 to debug the target executable.

   • For Linux: (gdb) Launch
   • For Windows: (Windows) Launch
   • For macOS: (lldb) Launch
     • You may need to enable Developer mode:

     $ sudo DevToolsSecurity -enable

Starting Your Project

1. Untrack the .vscode directory and commit the change.

   $ git rm --cached -r .vscode
   $ git commit -m "Remove VSCode settings"

2. Remove unnecessary directories and add_subdirectory entries in CMakeLists.txt.

3. Update project names and directories in CMakeLists.txt and vcpkg.json.

4. Add your dependencies to CMakeLists.txt and vcpkg.json.

Updating vcpkg

1. Update the vcpkg submodule to the specified commit hash.
2. Edit the builtin-baseline in vcpkg.json to match the hash above.