Struct AdjacencyGpu 

pub struct AdjacencyGpu {
    pub buffer: Buffer,
    pub max_neighbors: u32,
    pub radius: f32,
    /* private fields */
}

GPU resources for adjacency buffer management.

Fields

buffer: Buffer

The adjacency buffer. Layout per particle: [count: u32, neighbor_indices: array<u32, max_neighbors>] Total stride: (1 + max_neighbors) * 4 bytes

max_neighbors: u32

Maximum neighbors per particle.

radius: f32

Radius for neighbor detection.

Implementations

impl AdjacencyGpu

pub fn new( device: &Device, particle_buffer: &Buffer, spatial: &SpatialGpu, num_particles: u32, max_neighbors: u32, radius: f32, particle_stride: usize, ) -> Self

Create a new adjacency buffer system.

Arguments

• device - The wgpu device
• particle_buffer - Buffer containing particle data
• spatial - Spatial hashing system for neighbor queries
• num_particles - Number of particles
• max_neighbors - Maximum neighbors to store per particle
• radius - Radius for neighbor detection
• particle_stride - Byte stride of each particle in the buffer

pub fn execute(&self, encoder: &mut CommandEncoder)

Execute the adjacency computation pass.

This should be called after spatial hashing is executed but before the main particle compute shader.

pub fn stride_u32(&self) -> u32

Get the stride (in u32s) for each particle’s adjacency data.