pub fn cpu_parallel_dispatch<F>( workgroup_size: [u32; 3], groups: [u32; 3], num_threads: usize, f: F, )where F: Fn(usize, u32, u32, u32) + Send + Sync,
Run a CPU compute kernel across multiple threads. f receives (thread_id, global_x, global_y, global_z).
f
(thread_id, global_x, global_y, global_z)