webgpu-cross-platform-demo

This is a small test app that uses WebGPU’s webgpu.h header as a platform-agnostic hardware abstraction layer. It uses a C++ layer over webgpu.h, called webgpu_cpp.h.

On native platforms, this project can be built against Dawn, Chromium’s native WebGPU implementation.

On the Web, this project can be built against Emscripten, using Dawn’s “emdawnwebgpu” implementation of webgpu.h on top of the browser’s own WebGPU JavaScript API.

It currently hasn’t been set up to build against wgpu-native’s webgpu.h implementation, but that is a goal.

Check the issues tab for known issues.

Please note the webgpu.h API (implemented by Dawn/wgpu-native/Emscripten) and webgpu_cpp.h bindings (from Dawn) are not yet stabilized.

Open pre-installed VSCode in Browser

Open in Gitpod

Building

Instructions are for Linux/Mac; Windows builds should work, but you need to adjust the command syntax accordingly.

This setup step is always required for native builds. For web builds, it is only strictly needed if you need to rebuild emdawnwebgpu_pkg (but without it, you’ll need to build manually with your own emsdk, because build_all.sh uses Dawn’s).

Web build

The web build has been mainly tested with Chrome Canary on Mac, but it should work on any conformant browser and supported platform.

Once you have an Emscripten SDK, you can simply build it using emcmake. This will use the snapshot of emdawnwebgpu in this repository. You can also get newer releases of emdawnwebgpu from https://github.com/google/dawn/releases.

mkdir out/web
cd out/web
path/to/emcmake cmake ../.. -DDEMO_USE_ASYNCIFY=1  # Add other flags here as desired
make -j4

For development of emdawnwebgpu, there’s a script that will build a bunch of different variants. To use this, set up the Dawn dependency, which regenerates emdawnwebgpu_pkg, and then run this script (which uses Dawn’s copy of emsdk):

./setup_build.sh --checkout=1
./build_web_many_configs.sh --parallel=0

Native build

Build has been tested on Linux/Mac/Win10 (though may be broken on some platforms).

mkdir -p out/native
cd out/native

# Using Make:
cmake ../..  # Add -DCMAKE_BUILD_TYPE=Release, -DDEMO_USE_GLFW=OFF, etc. here as desired
make clean
make -j4 all

# Or, to use Ninja instead of Make:
cmake -GNinja ../..  # Add flags here as desired
ninja

Note: If you want to have window displayed, make sure to have glfw available.

How to config glfw package on each platform? - Linux ```sh apt-get install libglfw3-dev ``` - macOS ```sh brew install glfw ``` - Win - Manually - Download glfw source package and windows pre-compiled binaries from [here](https://www.glfw.org/download) - Unzip (e.g. version 3.3.8) `glfw-3.3.8.zip` to `C:/Program Files (x86)/glfw/` - Unzip (e.g. version 3.3.8) `glfw-3.3.8.bin.WIN64.zip` and put `lib-vc2022` to `C:/Program Files (x86)/glfw/lib-vc2022` (assume you are using Visual Studio 2022 to build) - You are done. `cmake/modules/Findglfw3.cmake` is used to find them (You can edit the path it uses) - vcpkg - Here is a fork that uses vcpkg to manage glfw: [link](https://github.com/bitsauce/webgpu-cross-platform-demo)

Alternatively, you can disable using glfw and window display by

cmake ../.. -DDEMO_USE_GLFW=OFF