pub fn gpu_dispatch_compute(pipeline: u32, x: u32, y: u32, z: u32) -> bool
Submit a compute dispatch with the given workgroup counts.