* wip

* add ref_by_trainer_id op

* ready to test

* fix ref inputs

* refine rpc_op_handle

* fix merge bug

* make deadline configurable

* configurable deadline

* update

* fix grpc deadline exceeded

The status is updated in the Process() thread
and can be checked in another HandleRequest() thread.

    Experiment on vgg flower, 2 trainers, 1ps.
    more trainer could have more speedup.

    After:
    Pass = 0, Iters = 327, Speed = (7.52) img/s
    Before:
    Pass = 0, Iters = 385, Speed = (6.77) img/s

* listen_and_serv use localscope

* fix ut