27e7bfe998
* Extreme WIP: New wgpu access api * Add getters for device and queue to Pixels * Don't run ScalingRenderer in render_custom() * Expose Pixels::scaling_renderer * Fix exposing ScalingRenderer * Tweak Pixels::render_custom() * Cleanup * More cleanup * Fix doc comment * Clippy * Fix doctests Co-authored-by: JMS55 <47158642+JMS55@users.noreply.github.com> |
||
---|---|---|
.github | ||
examples | ||
img | ||
pixels-mocks | ||
shaders | ||
src | ||
.gitignore | ||
.travis.yml | ||
Cargo.toml | ||
LICENSE | ||
README.md |
A tiny hardware-accelerated pixel frame buffer. 🦀
But why?
Rapidly prototype a simple 2D game, pixel-based animations, software renderers, or an emulator for your favorite platform. Then add shaders to simulate a CRT or just to spice it up with some nice VFX.
pixels
is more than just a library to push pixels to a screen, but less than a full framework. You're in charge of managing a window environment, event loop, and input handling.
Features
- Built on modern graphics APIs powered by
wgpu
: DirectX 12, Vulkan, Metal, OpenGL. - Use your own custom shaders for special effects. (WIP)
- Hardware accelerated scaling on perfect pixel boundaries.
- Supports non-square pixel aspect ratios. (WIP)
Examples
Troubleshooting
The most common issue is having an outdated graphics driver installed on the host machine. pixels
requests a low power (aka integrated) GPU by default. If the examples are not working for any reason, you may try setting the PIXELS_HIGH_PERF
environment variable (the value does not matter, e.g. PIXELS_HIGH_PERF=1
is fine) to see if that addresses the issue on your host machine.
You should also try to keep your graphics drivers up-to-date, especially if you have an old Intel integrated GPU. Keep in mind that some drivers and GPUs are EOL and will not be supported.
Comparison with minifb
The minifb
crate shares some similarities with pixels
; it also allows rapid prototyping of 2D games and emulators. But it requires the use of its own window/GUI management, event loop, and input handling. One of the disadvantages with the minifb
approach is the lack of hardware acceleration (except on macOS, which uses Metal but is not configurable). An advantage is that it relies on fewer dependencies.