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

Disable unstable unittest

remove unused nccl.cmake

[Feature] Enable multi gpu distributed training of fluid

Fix api docs display error for fluid Initializer

 Add contents for manully build documentation(cn version)

* "add metrics"

* "add fluid metrics"

* "add import guards"

* "show warnings"

* "add demo"

* "fix ci"

* "add some details"

* "fix cci"

* "add demo Python"

* "add metrics"

Modify readers to fit the parallel executor

Make cuda_helper.h Pass cpplint

k8s dist train for en

add remove_op, remove_var in Python end

* fix gemm error for cuda 7.5

* fix version number

Fix lost of LoD while splitting tensor in parallel executor.

* Update compile_paddle_lib_en.md

Fix https://github.com/PaddlePaddle/Paddle/issues/8916

* Update compile_paddle_lib_en.md