Communication optimizations for global multi-threaded instruction scheduling