wgpu
is a cross-platform, safe, pure-rust graphics api. It runs natively on Vulkan, Metal, D3D12, D3D11, and OpenGLES; and on top of WebGPU on wasm.
The api is based on the WebGPU standard. It serves as the core of the WebGPU integration in Firefox, Servo, and Deno.
The repository hosts the following libraries:
- User facing Rust API.
- Internal WebGPU implementation.
- Internal unsafe GPU API abstraction layer.
- Rust types shared between all crates.
- WebGPU implementation for the Deno JavaScript/TypeScript runtime
The following binaries:
cts_runner
- WebGPU Conformance Test Suite runner usingdeno_webgpu
.player
- standalone application for replaying the API traces.wgpu-info
- program that prints out information about all the adapters on the system or invokes a command for every adapter.
For an overview of all the components in the gfx-rs ecosystem, see the big picture.
Minimum Supported Rust Version is 1.64.
It is enforced on CI (in "/.github/workflows/ci.yml") with RUST_VERSION
variable.
This version can only be upgraded in breaking releases.
The wgpu-core
, wgpu-hal
, and wgpu-types
crates should never
require an MSRV ahead of Firefox's MSRV for nightly builds, as
determined by the value of MINIMUM_RUST_VERSION
in
python/mozboot/mozboot/util.py
. However, Firefox uses cargo vendor
to extract only those crates it actually uses, so the
workspace's other crates can have more recent MSRVs.
Note for Rust 1.64: The workspace itself can even use a newer MSRV
than Firefox, as long as the vendoring step's Cargo.toml
rewriting
removes any features Firefox's MSRV couldn't handle. For example,
wgpu
can use manifest key inheritance, added in Rust 1.64, even
before Firefox reaches that MSRV, because cargo vendor
copies
inherited values directly into the individual crates' Cargo.toml
files, producing 1.63-compatible files.
Rust examples can be found at wgpu/examples
. You can run the examples with cargo run --example name
. See the list of examples. For detailed instructions, look at Running the examples on the wiki.
If you are looking for a wgpu tutorial, look at the following:
To use wgpu in C/C++, you need wgpu-native.
If you want to use wgpu in other languages, there are many bindings to wgpu-native from languages such as Python, D, Julia, Kotlin, and more. See the list.
We have the Matrix space with a few different rooms that form the wgpu community:
- discussion of the library's development.
- discussion of using the library and the surrounding ecosystem.
- discussion of everything else.
We have a wiki that serves as a knowledge base.
API | Windows | Linux & Android | macOS & iOS |
---|---|---|---|
Vulkan | ✅ | ✅ | 🆗 (vulkan-portability) |
Metal | ✅ | ||
DX12 | ✅ (W10+ only) | ||
DX11 | 🛠️ | ||
GLES3 | 🆗 | ||
Angle | 🆗 | 🆗 | 🆗 (macOS only) |
✅ = First Class Support — 🆗 = Best Effort Support — 🛠️ = Unsupported, but support in progress
wgpu supports shaders in WGSL, SPIR-V, and GLSL. Both HLSL and GLSL have compilers to target SPIR-V. All of these shader languages can be used with any backend, we will handle all of the conversion. Additionally, support for these shader inputs is not going away.
While WebGPU does not support any shading language other than WGSL, we will automatically convert your non-WGSL shaders if you're running on WebGPU.
WGSL is always supported by default, but GLSL and SPIR-V need features enabled to compile in support.
Note that the WGSL specification is still under development,
so the draft specification does not exactly describe what wgpu
supports.
See below for details.
To enable SPIR-V shaders, enable the spirv
feature of wgpu.
To enable GLSL shaders, enable the glsl
feature of wgpu.
Angle is a translation layer from GLES to other backends, developed by Google. We support running our GLES3 backend over it in order to reach platforms with GLES2 or DX11 support, which aren't accessible otherwise. In order to run with Angle, "angle" feature has to be enabled, and Angle libraries placed in a location visible to the application. These binaries can be downloaded from gfbuild-angle artifacts, manual compilation may be required on Macs with Apple silicon.
On Windows, you generally need to copy them into the working directory, in the same directory as the executable, or somewhere in your path.
On Linux, you can point to them using LD_LIBRARY_PATH
environment.
All testing and example infrastructure shares the same set of environment variables that determine which Backend/GPU it will run on.
WGPU_ADAPTER_NAME
with a substring of the name of the adapter you want to use (ex.1080
will matchNVIDIA GeForce 1080ti
).WGPU_BACKEND
with a comma separated list of the backends you want to use (vulkan
,metal
,dx12
,dx11
, orgl
).WGPU_POWER_PREF
with the power preference to choose when a specific adapter name isn't specified (high
orlow
)WGPU_DX12_COMPILER
with the DX12 shader compiler you wish to use (dxc
orfxc
, note thatdxc
requiresdxil.dll
anddxcompiler.dll
to be in the working directory otherwise it will fall back tofxc
)
When running the CTS, use the variables DENO_WEBGPU_ADAPTER_NAME
, DENO_WEBGPU_BACKEND
, DENO_WEBGPU_POWER_PREFERENCE
.
We have multiple methods of testing, each of which tests different qualities about wgpu. We automatically run our tests on CI if possible. The current state of CI testing:
Backend/Platform | Tests | CTS | Notes |
---|---|---|---|
DX12/Windows 10 | ✔️ | ✔️ | using WARP |
DX11/Windows 10 | 🚧 | — | using WARP |
Metal/MacOS | — | — | metal requires GPU |
Vulkan/Linux | ✔️ | ❌ | using lavapipe, cts hangs |
GLES/Linux | ✔️ | — | using llvmpipe |
We use a tool called cargo nextest
to run our tests.
To install it, run cargo install cargo-nextest
.
To run the test suite on the default device:
cargo nextest run --no-fail-fast
wgpu-info
can run the tests once for each adapter on your system.
cargo run --bin wgpu-info -- cargo nextest run --no-fail-fast
Then to run an example's image comparison tests, run:
cargo nextest run <example-test-name> --no-fail-fast
Or run a part of the integration test suite:
cargo nextest run -p wgpu -- <name-of-test>
If you are a user and want a way to help contribute to wgpu, we always need more help writing test cases.
WebGPU includes a Conformance Test Suite to validate that implementations are working correctly. We can run this CTS against wgpu.
To run the CTS, first you need to check it out:
git clone https://github.com/gpuweb/cts.git
cd cts
# works in bash and powershell
git checkout $(cat ../cts_runner/revision.txt)
To run a given set of tests:
# Must be inside the cts folder we just checked out, else this will fail
cargo run --manifest-path ../cts_runner/Cargo.toml -- ./tools/run_deno --verbose "<test string>"
To find the full list of tests, go to the online cts viewer.
The list of currently enabled CTS tests can be found here.
The wgpu
crate is meant to be an idiomatic Rust translation of the WebGPU API.
That specification, along with its shading language, WGSL,
are both still in the "Working Draft" phase,
and while the general outlines are stable,
details change frequently.
Until the specification is stabilized, the wgpu
crate and the version of WGSL it implements
will likely differ from what is specified,
as the implementation catches up.
Exactly which WGSL features wgpu
supports depends on how you are using it:
-
When running as native code,
wgpu
uses the Naga crate to translate WGSL code into the shading language of your platform's native GPU API. Naga has a milestone for catching up to the WGSL specification, but in general there is no up-to-date summary of the differences between Naga and the WGSL spec. -
When running in a web browser (by compilation to WebAssembly) without the
"webgl"
feature enabled,wgpu
relies on the browser's own WebGPU implementation. WGSL shaders are simply passed through to the browser, so that determines which WGSL features you can use. -
When running in a web browser with
wgpu
's"webgl"
feature enabled,wgpu
uses Naga to translate WGSL programs into GLSL. This uses the same version of Naga as if you were runningwgpu
as native code.
wgpu uses the coordinate systems of D3D and Metal:
Render | Texture |
---|---|
![]() |
![]() |