You can adjust process priorities, which influence how the cores pick work from a shared work queue - and we preempt processes that do not block to balance the work across the cores.