Table of Contents:

• Prerequisites
• Generating Rust glue
• Generating static library
• Running an example

Before getting started with wgpu, there are some necessary components that have to be installed and configured.

To build wgpu-native:

• Rust compiler
• LLVM with CLang
  • On MacOS: No action required.
  • On Linux: E.g. yum install clang.
  • On Windows, LLVM can be installed via the binary release or using Chocolatey. Then set the LIBCLANG_PATH (default location is C:\Program Files\LLVM\lib (or C:\Program Files (x86)\LLVM\lib for x86).
• (Windows only) Updated Windows SDK

To build native examples:

• C compiler which matches the Rust target (e.g. MSVC, Clang, or GCC)
• GLFW 3
• (Windows only) MSBuild

To ensure the examples run correctly, it's also useful to ensure drivers for any installed graphics APIs (e.g. Vulkan, DirectX 12, etc.) are up-to-date.

You can clone the wgpu-native repo as you'd normally do, but you also need to initialize the submodules to populate the ffi/webgpu-headers directory:

git submodule update --init

You may need to do this once in a while:

rustup update
cargo update

TODO

Rust code can be compiled separately into a static library.

In the Makefile, the static library commands are prefixed by lib-.

To generate the wgpu static library, use the following command:

make lib-native

The static libraries will be located in the target folder in the root of the project.

The examples are located under the examples directory. We are using a Makefile for running the examples.

In the Makefile, the example commands are prefixed by example-.

To generate the Triangle example, use the following command:

make run-example-triangle

To generate the Compute example, use the following command:

make run-example-compute

To generate the Capture example, use the following command:

make run-example-capture