Improve bulk implementation for static_thread_pool

[maikel]

The current implementation of bulk doesn't use the load balancer facility and should probably schedule chunked index ranges. The optimal chunk size is application specific, so one might add a way specify that somehow