Struct CpuComputeDispatch 

pub struct CpuComputeDispatch {
    pub kernel: ComputeKernelKind,
    pub workgroup_size: WorkgroupSize,
}

Dispatcher that executes compute kernels on the CPU.

Fields

kernel: ComputeKernelKind

Which kernel this dispatcher is configured for.

workgroup_size: WorkgroupSize

Workgroup size hint (informational on the CPU path).

Implementations

impl CpuComputeDispatch

pub fn new(kernel: ComputeKernelKind, wg: WorkgroupSize) -> Self

Create a new dispatcher.

pub fn dispatch_velocity_update( &self, pos: &mut ComputeBuffer, vel: &mut ComputeBuffer, force: &ComputeBuffer, mass: &ComputeBuffer, dt: f32, n: usize, )

Semi-implicit Euler integration for n particles.

pos[i] += vel[i] * dt then vel[i] += force[i] / mass[i] * dt.

pub fn dispatch_pressure_jacobi( &self, p: &mut ComputeBuffer, p_old: &ComputeBuffer, rhs: &ComputeBuffer, nx: usize, ny: usize, dx: f32, )

Single Jacobi pressure iteration over an nx × ny grid.

Interior points only; boundary cells are left unchanged.

p[i,j] = (p_old[i+1,j] + p_old[i-1,j] + p_old[i,j+1] + p_old[i,j-1] - dx² * rhs[i,j]) / 4

pub fn dispatch_particle_force( &self, pos: &ComputeBuffer, force: &mut ComputeBuffer, eps: f32, sigma: f32, n: usize, )

O(n²) Lennard-Jones force accumulation.

Positions are stored as interleaved [x0, y0, x1, y1, …]. Forces are accumulated in-place (force is zeroed first).