Struct GpuMaterial 

pub struct GpuMaterial {

    pub uniform_buffer: BufferId,
    pub base_color_view: Option<TextureViewId>,
    pub metallic_roughness_view: Option<TextureViewId>,
    pub normal_view: Option<TextureViewId>,
    pub emissive_view: Option<TextureViewId>,
    pub occlusion_view: Option<TextureViewId>,
    pub base_color_texture: Option<TextureId>,
    pub metallic_roughness_texture: Option<TextureId>,
    pub normal_texture: Option<TextureId>,
    pub emissive_texture: Option<TextureId>,
    pub occlusion_texture: Option<TextureId>,
    pub sampler: SamplerId,
    pub bind_group: BindGroupId,
    pub variant: ShaderVariantKey,
    pub double_sided: bool,
    pub blend: bool,
}

A GPU-ready representation of a material: its uniform buffer, the PBR texture views it declares, the shared sampler, and a ready-to-bind group-2 bind group.

Built once per material UUID by the data-layer projection (mirroring GpuMesh), cached, and consumed by every lit lane — which selects the pipeline for variant and binds bind_group at group 2. Each texture view is Some only for a map the material declares (no fallback textures); the variant records exactly that set of HAS_* flags, so the bind group, the group-2 layout it was built against, and the lit pipeline’s group-2 layout all resolve from the same (LayoutKey::Material, variant).

Each declared map is uploaded as a texture private to this material (the projection does not share GPU textures across materials), so the material exclusively owns its uniform_buffer, its four *_texture / *_view handles, and its bind_group. The sampler is the opposite: it is the engine-shared filtering sampler. This ownership split is what asset eviction relies on to free the resources — it destroys everything but the shared sampler when the material is reclaimed. The *_texture and matching *_view fields are always Some/None together.

Fields

uniform_buffer: BufferId

MaterialUniforms uniform buffer (base color, factors, …).

base_color_view: Option<TextureViewId>

Base-color (albedo) texture view, if declared.

metallic_roughness_view: Option<TextureViewId>

Metallic-roughness texture view (B=metallic, G=roughness), if declared.

normal_view: Option<TextureViewId>

Tangent-space normal texture view, if declared.

emissive_view: Option<TextureViewId>

Emissive texture view, if declared.

occlusion_view: Option<TextureViewId>

Ambient-occlusion texture view, if declared.

base_color_texture: Option<TextureId>

Base-color texture backing base_color_view, owned by this material and freed on eviction.

metallic_roughness_texture: Option<TextureId>

Metallic-roughness texture backing metallic_roughness_view.

normal_texture: Option<TextureId>

Normal texture backing normal_view.

emissive_texture: Option<TextureId>

Emissive texture backing emissive_view.

occlusion_texture: Option<TextureId>

Ambient-occlusion texture backing occlusion_view.

sampler: SamplerId

Filtering sampler shared by all maps (engine-owned; NOT freed on eviction).

bind_group: BindGroupId

Prebuilt group-2 bind group bound by lit lanes.

variant: ShaderVariantKey

Shader variant (HAS_* texture flags) this material was built for. The lit lane requests the matching pipeline and groups draws by it.

double_sided: bool

Whether the material renders double-sided (back faces not culled). The lit lane selects a no-cull pipeline for it; single-sided materials cull back faces. Part of the pipeline cache key via [PipelineKey].

blend: bool

Whether the material is alpha-blended (AlphaMode::Blend).

Blended materials need a different pipeline state (alpha blending on, depth writes off) and a different draw order: the lit lane defers them to a second, back-to-front sorted batch after the opaque draws, because blending is order-dependent. AlphaMode::Mask is not blended — it discards in the shader and stays in the opaque batch.

Implementations

impl GpuMaterial

pub fn bindings(&self) -> MaterialGpuBindings

Returns the [MaterialGpuBindings] view of this material’s resources, suitable for fill_material_bind_group_entries.

Trait Implementations

impl Clone for GpuMaterial

fn clone(&self) -> GpuMaterial

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for GpuMaterial

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Asset for GpuMaterial