Struct CollisionGpuPipeline 

pub struct CollisionGpuPipeline {
    pub broadphase: BroadphaseGpuKernel,
    pub narrowphase: NarrowphaseGpuKernel,
    pub manifolds: Vec<PersistentManifoldGpu>,
    pub contacts: Vec<ContactResult>,
}

Combined broadphase + narrowphase collision pipeline.

Holds the configuration for both stages and maintains a list of persistent manifolds from the previous frame for warm starting.

Fields

broadphase: BroadphaseGpuKernel

Broadphase kernel configuration.

narrowphase: NarrowphaseGpuKernel

Narrowphase kernel.

manifolds: Vec<PersistentManifoldGpu>

Persistent manifolds from the previous timestep.

contacts: Vec<ContactResult>

Contact list from the most recent dispatch.

Implementations

impl CollisionGpuPipeline

pub fn new(margin: f32) -> Self

Create a new pipeline with default broadphase margin.

pub fn dispatch(&mut self, aabbs: &[AabbGpu])

Run one full collision detection pass over aabbs.

1. Broadphase: find overlapping AABB pairs.
2. Narrowphase: compute contact points for each pair.
3. Update persistent manifolds.

pub fn dispatch_bvh(&mut self, aabbs: &[AabbGpu])

Run the full pipeline using BVH-accelerated broadphase.

pub fn reset(&mut self)

Clear all manifolds and contact data.

pub fn total_contact_points(&self) -> usize

Return the total number of contact points across all manifolds.

Trait Implementations

impl Debug for CollisionGpuPipeline

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

Formats the value using the given formatter.