Expand description
GPU-accelerated collision detection (CPU mock backend via Rayon).
Implements a broadphase AABB sweep-and-prune followed by a narrowphase sphere-sphere test and a stub GJK dispatcher. The “GPU” dispatch is mocked using Rayon parallel iterators so the module compiles and runs on CPU.
Structs§
- Aabb
- Axis-aligned bounding box.
- GpuCollision
Buffer - Buffer holding the AABB list and broadphase potential pairs.
- Sphere
- Sphere descriptor: centre + radius.
- Sphere
Contact - Sphere-sphere narrowphase collision result.
Enums§
- GjkResult
- Stub result from the GJK narrowphase.
Functions§
- batch_
gjk_ dispatch - Dispatch GJK for a batch of collision pairs (stub implementation).
- build_
collision_ pairs - Build the full broadphase collision pair list from a sorted AABB list.
- gpu_
aabb_ overlap - AABB-AABB overlap test kernel (mock GPU dispatch).
- gpu_
broadphase_ sort - Sort AABBs by their minimum x-coordinate (mock GPU sort).
- gpu_
sphere_ collision - Test sphere-sphere collisions for the given pairs (mock GPU narrowphase).