wgpu-rs is an idiomatic Rust wrapper over wgpu-core. It's designed to be suitable for general purpose graphics and computation needs of Rust community.
wgpu-rs can target both the natively supported backends and WASM directly.
See our gallery and the wiki page for the list of libraries and applications using wgpu-rs
.
All examples are located under the examples directory.
These examples use the default syntax for running examples, as found in the Cargo documentation. For example, to run the cube
example:
cargo run --example cube
The hello-triangle
and hello-compute
examples show bare-bones setup without any helper code. For hello-compute
, pass 4 numbers separated by spaces as arguments:
cargo run --example hello-compute 1 2 3 4
Running on the web is still work-in-progress. You may need to enable experimental flags on your browser. Check browser implementation status on webgpu.io. Notably, wgpu-rs
is often ahead in catching up with upstream WebGPU API changes. We keep the gecko
branch pointing to the code that should work on latest Firefox.
To run examples on the wasm32-unknown-unknown
target, first build the example as usual, then run wasm-bindgen
:
# Checkout `gecko` branch that matches the state of Firefox
git checkout upstream/gecko
# Install or update wasm-bindgen-cli
cargo install -f wasm-bindgen-cli
# Build with the wasm target
RUSTFLAGS=--cfg=web_sys_unstable_apis cargo build --target wasm32-unknown-unknown --example hello-triangle
# Generate bindings in a `target/generated` directory
wasm-bindgen --out-dir target/generated --web target/wasm32-unknown-unknown/debug/examples/hello-triangle.wasm
Create an index.html
file into target/generated
directory and add the following code:
<html>
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
</head>
<body>
<script type="module">
import init from "./hello-triangle.js";
init();
</script>
</body>
</html>
Now run a web server locally inside the target/generated
directory to see the hello-triangle
in the browser.
e.g. python -m http.server
Currently, shaders in the examples are written in GLSL 4.50 and compiled to SPIR-V manually. In the future WGSL will be the shader language for WebGPU, but support is not implemented yet.
For now, the shaders can be compiled to SPIR-V by running make
, which requires you to have glslang
s glslangValidator
binary.
If you need to test local fixes to gfx-rs or other dependencies, the simplest way is to add a Cargo patch. For example, when working on DX12 backend on Windows, you can check out the "hal-0.2" branch of gfx-rs repo and add this to the end of "Cargo.toml":
[patch.crates-io]
gfx-backend-dx12 = { path = "../gfx/src/backend/dx12" }
gfx-hal = { path = "../gfx/src/hal" }
If a version needs to be changed, you need to do cargo update -p gfx-backend-dx12
.