Function fdtd_tile_step_dsl 

pub fn fdtd_tile_step_dsl(
    pressure: &[f32],
    pressure_prev: &mut [f32],
    c2: f32,
    damping: f32,
    tile_size: usize,
    halo: usize,
)

Run FDTD on an entire tile using the DSL kernel.

This simulates what CUDA would do: run the kernel for each interior cell.