fix LogActivation is not defined

Refine hierarchical-layer.rst

Add GPU Profiler in PaddlePaddle

merge Release/v0.9.0 to develop

merge bugfix #593 and # 597 from develop branch

Revise doc and some scripts for demo.

Add PaddlePaddle QuickStart Demo on Kubernetes

revise input order and sort bug

fix the has_LayerOutput boolean for StaticInput as the input of recurrent_group

Fix a wrong link in build doc

 fix documentation

refine compile_options.rst

Fix the config_parse.py if user does not set padding for Pool in the old config.

clang format .cc .h .cpp .c and .hpp file

Feature/add clang format plugin

Fix badge's wrong link in Readme

Change demo datafile location by using CDN in baidu.

* Merge conflict using theirs

* Because it is a cuda source file, and we need to support c++ 03 in
  cuda.

* Different Type compare.
* ostream << should pass a const object.
* remove always true checks.