# Custom GPU Resources
Anything one-off that needs the device before it can be built and isn't a material/mesh/texture — a camera, a depth buffer — is a [`LazyResource`](./the-asset-pipeline.md#lazyresourceb-exactly-one-constructed-on-demand), constructed once `WGPUBackend` (or your own `Deps`) exists, built entirely from the opaque builders covered elsewhere in this book. There's no built-in `pebble::wgpu` camera type — a camera's uniform layout is yours to define — so this page walks through building one by hand.
## The depth texture
A depth buffer has no source data to upload — it's not what [`Texture`](./textures.md) is for (that loads pixel data from a file/bytes). [`TextureBuilder`](./textures.md#a-render-target--depth-buffer-no-source-data) is the builder for exactly this: an empty GPU-side texture, handed back as an opaque `TextureView` ready to use as a render target:
```rust
use pebble::wgpu::prelude::*;
struct DepthTexture {
view: TextureView,
}
impl LazyResource<WGPUBackend> for DepthTexture {
type Deps<'a> = ();
fn construct<'a>(backend: &WGPUBackend, _deps: &()) -> Option<Self> {
let view = TextureBuilder::new(backend.surface_width(), backend.surface_height(), TextureFormat::Depth16Unorm)
.label("depth")
.usage(TextureUsages::RENDER_ATTACHMENT)
.build(backend);
Some(DepthTexture { view })
}
}
```
```rust
.add_plugin(LazyResourcePlugin::<WGPUBackend, DepthTexture>::new())
```
## The camera
A camera needs a uniform buffer (the view/projection matrices), a bind group layout describing that buffer, and a bind group binding the two together — all built once the device exists. `wgpu::prelude` (imported above, alongside `WGPUBackend`) is where the builders below live — `BindGroupLayoutBuilder`, `BufferBuilder`, `BindGroupBuilder` — reach for those over hand-writing a `wgpu::BufferDescriptor`/`BindGroupLayoutDescriptor`/`BindGroupDescriptor` by hand. Every value that comes back — `BindGroupLayout`, `Buffer`, `BindGroup` — is opaque, the same as everywhere else in `pebble::wgpu`: no `wgpu::*` type anywhere in `Camera`'s own definition.
```rust
struct Camera {
buffer: Buffer,
bind_group_layout: BindGroupLayout,
bind_group: BindGroup,
}
impl LazyResource<WGPUBackend> for Camera {
type Deps<'a> = ();
fn construct<'a>(backend: &WGPUBackend, _deps: &()) -> Option<Self> {
// Same BindGroupLayoutBuilder that Material/Compute use internally
// (see Materials and Compute Pipelines) — a camera's
// layout isn't going through build_material, but there's no reason to
// hand-write a wgpu::BindGroupLayoutDescriptor when the same builder
// already covers a single uniform-buffer entry.
let bind_group_layout = BindGroupLayoutBuilder::new()
.label("camera_layout")
.entry("camera", 0, BindingKind::uniform_buffer(ShaderStages::VERTEX))
.build(backend);
// Empty for now — there's no view/projection data yet to seed it
// with; written every frame via Buffer::write once the actual
// matrices are known (see below).
let size = std::mem::size_of::<[[f32; 4]; 4]>() as u64 * 2; // view + projection
let buffer = BufferBuilder::empty(size).label("camera").uniform().build(backend);
let bind_group = BindGroupBuilder::new(&bind_group_layout)
.label("camera_bind_group")
.buffer(&buffer)
.build(backend);
Some(Camera { buffer, bind_group_layout, bind_group })
}
}
```
```rust
.add_plugin(LazyResourcePlugin::<WGPUBackend, Camera>::new())
```
Updating it every frame is an ordinary `Update`-stage system, writing fresh matrices via `camera.buffer.write(&bytes)` — nothing new relative to the [buffer basics](./buffers.md).
## Wiring the camera into a material's pipeline layout
A material's own texture/sampler bind group occupies `@group(0)`. The camera needs its own group too — `.entries(...)` (see [Bind Groups and Layouts](./bind-groups.md#pipeline-layouts-multiple-bind-groups)) takes a `GroupEntry::Layout` for exactly this: a bind group layout that exists *outside* a material's own entries, in whatever position matches the shader's `@group(N)`:
```rust
use pebble::wgpu::layout::GroupEntry;
fn setup(
// ...
camera: Res<Camera>,
depth: Res<DepthTexture>,
) -> Option<()> {
let material = Material::new(SHADER)
// ... label, vertex_layouts as usual ...
.entries(vec![
GroupEntry::Own(material_entries()), // @group(0): albedo/sampler
GroupEntry::Layout(camera.bind_group_layout.clone()), // @group(1): camera
])
.depth(DepthStencilState {
format: TextureFormat::Depth16Unorm,
depth_write_enabled: Some(true),
depth_compare: Some(CompareFunction::Less),
stencil: StencilState::default(),
bias: DepthBiasState::default(),
})
.build_asset("lit", &mut materials);
// ...
Some(())
}
```
`GroupEntry::Layout` takes the same opaque `BindGroupLayout` `bind_group_layout` above — `Clone` because a camera's layout might be wired into more than one material, and it's the same cheap `Arc`-backed handle underneath either way. Position in `.entries(...)` *is* the `@group(N)` index, so `camera.bind_group_layout` landing at `@group(1)` here is just "it's the second element" — no separate group number to keep in sync with the shader by hand. `build_material` still panics on more than one `GroupEntry::Own` or on exceeding the device's `max_bind_groups`, turning either mistake into an immediate, specific error instead of an opaque wgpu validation failure at draw time.
`setup` requiring `Res<Camera>`/`Res<DepthTexture>` (hard requirements, see [Resources](./resources.md#what-happens-when-a-resource-isnt-there-yet)) is what makes this correct without any manual waiting: `setup` itself won't run until both lazy resources exist, so by the time it builds the material, `camera.bind_group_layout` is guaranteed to be real.
A camera shared across *every* material, rather than wired in by hand at each call site, is exactly what [`GlobalLayoutPool`](./bind-groups.md#a-pool-of-shared-layouts) is for — register `camera.bind_group_layout` into it once, and pull it with `pool.get("camera")` wherever it's needed.
## Rendering with a depth attachment
`render_context` is a shortcut for "one color attachment, no depth." A depth pass uses `begin_pass` directly — see [Recording a Render Pass](./rendering-pass-recording.md#a-custom-render-target-or-depth-attachment):
```rust
use pebble::prelude::{ColorTarget, DepthTarget, Pass};
fn render(
mut frame: ResMut<CurrentFrame<WGPUBackend>>,
camera: Res<Camera>,
depth: Res<DepthTexture>,
// ... materials, meshes, instances as usual ...
) {
let Some(mut active) = frame.active() else { return };
let mut pass = active.begin_pass(Pass {
colors: &[ColorTarget::default([0.2, 0.3, 0.3, 1.0])],
depth: Some(DepthTarget::new(&depth.view, 1.0)),
});
pass.set_bind_group(1, &camera.bind_group, &[]); // group 1: shared across every draw
for /* ... */ {
pass.set_pipeline(&material.pipeline);
pass.set_bind_group(0, &instance.bind_group, &[]); // group 0: per-instance
// set_vertex_buffer / set_index_buffer / draw_indexed as usual
}
}
```
`DepthTarget::new(view, 1.0)` clears the depth buffer to the far plane (`1.0`) at the start of the pass — a fragment only writes if its depth compares `Less` than what's already there, so nearer geometry always wins regardless of draw order. Bind group 1 (the camera) is set once per pass, outside the loop, since every draw shares the same view/projection; bind group 0 (the material instance) is set per-draw, inside it.
This same pattern — a `LazyResource` wrapping opaque builders, wired into a material via `GroupEntry::Layout` — is how any one-off GPU resource gets built: a shadow-map pass's own uniform buffer, a global lighting bind group, anything that's "exactly one of, needs the device to exist first."