* Add sub_blocks of lr_decay_op to pserver_prog after distribute_transpiler

* Remove unused logs and logics

* 1. Add ops to new block (considering the nested block condition)
2. Follow the original hierarchy of blocks
3. Change the function's name and remove debug lines

* wip

* use executor.complete to end trainer

* fix build

* fix build with distribute off

* fix typo

* fix cmake typo

* fix build

    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

Add the call of DropKids at the end of executor.Run to delete the local scopes created in operators (#10403)

* Add the call of DeleteScope to delete the memory of scope created by NewScope.

* Call DropKids at the end of executor.Run to delete all local scopes created in operators.

Add an argument in Executor.Run to allow users to choose whether to create and destroy variables every time. (#9242)

optimize executor.run

Single device on se-resnet reduce from 0.56 to 0.50

* Adding set_capacity method support

* Adding Python for make_channel

* Updating notest_concurrency

* Write python for make_channel method

* Write python for make_channel method

* Fix make_channel and test

* Placeholder ops for channel send, recv and close

* Adding ToTypeIndex method to var_type.h

* Add var_type.h to channel:

* Added POD_Type to the method

* Add CHANNEL to executor

* Updated get and set DataType to accomodate Channels

* Updating get and set to incorporate channels

* Adding CHANNEL as supported VarType in protobuf

* Removing unecessary import

* Fixing VarDesc to adapt to Channel as VarType

* Add channel.h to executor

* Remove innclude from channel

* Updated var_type to support Channel as  var type

* Adding get_channel to pybind

* Added ChannelHolder

* Adding make_channel as an op

* Adding ChannelHolder in channel

* Fixing typo

* Commenting out operators in concurrency

* Removing totypeid right now since we don't need it.

* Reverting python changes

* Fixing typo in framework.py

* Modify comments for ReaderHolder

* Support parallel test for understand_sentiment

* Full test on understand_sentiment

* Skip normal tests

* Debug CI

* Enable benchmark

* Revert init.cc

* Make CI pass