git clone https://github.com/snowapril/vkm.git
cd vkm
python bootstrap.py
mkdir build
cd build
cmake .. && cmake --build .

• vulkan-hpp
• glfw
• glm
• imgui
• ImGuizmo
• stb
• taskflow
• doctest

After running bootstrap.py, use the platform-appropriate runner script.

macOS / Linux

# Default backend (Metal on macOS, Vulkan on Linux)
bash scripts/run_tests.sh

# Explicit backend
bash scripts/run_tests.sh --backend metal
bash scripts/run_tests.sh --backend vulkan
bash scripts/run_tests.sh --backend webgpu

# All supported backends on the current platform
bash scripts/run_tests.sh --backend all

# Release build
bash scripts/run_tests.sh --build-type Release

run_tests.sh uses associative arrays and needs bash 4+; macOS ships bash 3.2, so install a newer one first (e.g. brew install bash) and invoke it explicitly if bash --version shows 3.x ($(brew --prefix)/bin/bash scripts/run_tests.sh ...).

Windows (PowerShell)

.\scripts\run_tests.ps1
.\scripts\run_tests.ps1 -BuildType Release

Each backend is built in its own isolated directory (build/metal/, build/vulkan/, build/webgpu/). Pass --no-deps-check / -NoDepsCheck to skip the dependencies/src/ existence check if you have already run the bootstrap script.

The webgpu backend targets Emscripten/WASM and is executed headlessly in a local Chrome or Chromium install (Node.js can't run it — it lacks navigator.gpu). It requires the emsdk toolchain, which bootstrap.py installs into dependencies/src/emsdk; if either emsdk or Chrome/Chromium isn't found, it's skipped with an informational message rather than failing.

After running bootstrap.py, use scripts/run_sample.* to configure, build, and run any sample for a given backend in one step.

macOS / Linux

bash scripts/run_sample.sh --backend metal --sample triangle
bash scripts/run_sample.sh --backend vulkan --sample triangle
bash scripts/run_sample.sh --backend webgpu --sample triangle

Windows (PowerShell)

.\scripts\run_sample.ps1 -Backend vulkan -Sample triangle
.\scripts\run_sample.ps1 -Backend webgpu -Sample triangle

Native backends (metal, vulkan) build into build/<backend>/ and launch the sample directly as a normal GUI window. The webgpu backend builds into build-wasm/ (matching the manual WebGPU instructions below) and opens the sample in your default browser via a short-lived local HTTP server. See "Native Backends" and "WebGPU" below for the equivalent manual, step-by-step commands each of these scripts automates.

Metal and Vulkan are both built directly by CMake (no browser/emsdk step). Metal is available on Apple platforms only and requires the Metal 4 (MTL4*) headers, which ship starting with the macOS 26 SDK — older Xcode/SDK installs will not have them. Vulkan is available on Linux and Windows natively, and on macOS via MoltenVK; it requires a Vulkan loader/SDK to be discoverable (either the VULKAN_SDK environment variable set, or vulkaninfo on PATH) — install the LunarG Vulkan SDK if you don't already have one. No separate setup step is needed beyond python bootstrap.py.

Known limitation: building the triangle sample with the Vulkan backend on macOS currently fails to link. The Apple platform windowing/render-coordinator code (src/vkm/platform/apple/application.mm) references the Metal-specific driver/swapchain types directly, regardless of which backend is selected, so a Vulkan-only build never compiles those types in. This doesn't affect Vulkan on Linux/Windows, or the UnitTests target on macOS (it doesn't create a window). Use Metal for the sample on macOS until application.mm is updated to dispatch by backend.

Building the sample manually (Metal)

cmake -B build/metal -DCMAKE_BUILD_TYPE=Release \
  -DVKM_USE_METAL_API=ON -DVKM_USE_VULKAN_API=OFF -DBUILD_SAMPLES=ON
cmake --build build/metal --target triangle

Building the sample manually (Vulkan)

cmake -B build/vulkan -DCMAKE_BUILD_TYPE=Release \
  -DVKM_USE_VULKAN_API=ON -DVKM_USE_METAL_API=OFF -DBUILD_SAMPLES=ON
cmake --build build/vulkan --target triangle

Running it

./build/metal/bin/triangle    # Metal
./build/vulkan/bin/triangle   # Vulkan (triangle.exe on Windows)

For the fully-automated path, prefer scripts/run_sample.py --backend metal --sample triangle (or --backend vulkan, or run_sample.sh/run_sample.ps1, see "Running Samples" above) over the manual steps here.

python bootstrap.py already installs and pins the emsdk toolchain into dependencies/src/emsdk — no separate setup step needed. The only other prerequisite is a local Chrome or Chromium install.

Building the sample manually

source dependencies/src/emsdk/emsdk_env.sh   # emsdk_env.bat on Windows
emcmake cmake -B build-wasm -DCMAKE_BUILD_TYPE=Release
cmake --build build-wasm --target triangle

Running it in a browser — serve the output directory and open it in Chrome:

cd build-wasm/bin
python3 -m http.server 8080
# open http://localhost:8080/triangle.html in Chrome

A normal (non-headless) Chrome window has WebGPU available by default on supported platforms — the --enable-unsafe-webgpu --use-angle=swiftshader flags below are a headless/software-rendering concern, not something you need for a manual, GUI browser check.

Building/running the WebGPU unit tests manually — the same emcmake/cmake --build --target UnitTests sequence produces build-wasm/bin/UnitTests.html. For interactive debugging, just open it in a normal Chrome window and watch doctest's output in DevTools. To reproduce exactly what the automated runner (scripts/run_tests.py --backend webgpu, see above) does headlessly:

google-chrome --headless=new --disable-gpu-sandbox --enable-unsafe-webgpu \
  --enable-features=Vulkan --use-angle=swiftshader --enable-logging=stderr --v=1 \
  http://localhost:8080/UnitTests.html

(macOS: use /Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome in place of google-chrome.)

For the fully-automated path, prefer scripts/run_tests.py --backend webgpu (or run_tests.sh/run_tests.ps1, see "Running Unit Tests" above) over the manual steps here — it also handles the local HTTP server for you.

For the fully-automated sample path (as opposed to tests), use scripts/run_sample.py --backend webgpu --sample triangle (or run_sample.sh/run_sample.ps1, see "Running Samples" above) — it builds via emcmake, serves build-wasm/bin locally, and opens triangle.html directly in your default browser.

Contributions are always welcome, either reporting issues/bugs or forking the repository and then issuing pull requests when you have completed some additional coding that you feel will be beneficial to the main project. If you are interested in contributing in a more dedicated capacity, then please contact me.

You can contact me via e-mail (sinjihng at gmail.com). I am always happy to answer questions or help with any issues you might have, and please be sure to share any additional work or your creations with me, I love seeing what other people are making.

The class is licensed under the MIT License:

Copyright (c) 2024 Snowapril

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.