* Fix cpplint warnings in beam_search_decode_op

* Fix cpplint warnings in beam_search_op

* Fix test

* fix

* Fix conv_op.h

* Fix conv_mkldnn_op

* Fix cpplint issues in detection_map_op

Add dist unitest data compare, ensure that dist train have same behavior as local training

Regenerate initializer.rst and add MSRAInitializer to initializer.__all__

Mpi enabled

Using constructor for VarHandle

move paddle/scripts/check_env.sh to benchmark/

The mkldnn tests are skipped when the MKLDNN flag is OFF

auto find tensorrt library and install in user root

Feature/tuning pe trans

Simplify DelayOps Logic

specify version for pip

Specified pip version in the dockerfile of production image

* add asynchronous design doc

* update distributed_training.png

* add distributed_training.graffle

* add async_distributed_training.png

* update design doc

* change name to distributed_training_overview

* follow comment

* change distributed_training_overview.md to README.md

* update sync_distributed_training.png

* follow comment

* fix typo

Add async update design doc