* wip add se-resnext test for 2x2 and parallel exe

* testing

* finish dist unitest 2x2

* update python interp

* fix gpu number

* make this test can run in parallel

Change and polish readers

* Fix Reduce and Gather

* Fix unit test

Fix deserialize bug

add option split mkl gemm

Fix IR CMake dependence.

fix AUC op
optimize it's test

fix mac build

Update version via branch name

Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/combine_open_files_and_double_buffer

Correct IR dependencies

Get more accurate file changes

Update hierarchical_layer_en.rst

* Update index_en.rst

* Update index_en.rst

* Update index_en.rst

Fix default value

Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/combine_open_files_and_double_buffer

Initial IR change

Fixed unittests for WITH_GPU=OFF and WITH_DISTRIBUTE=OFF build

fix dead link in high_level_api.md

better installation doc

fix mac build

* Added beam search API to fluid/contrib

* Change as review comments

* API.spec change

* Doc updated

Change the default value of the parameter 'drop_last' in 'paddle.batch' to False.

Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into feature/combine_open_files_and_double_buffer

fix _clone_variable in optimizer.py