* Add profiling information for inference example, recognize digits.

* Refine the profiling method.

* Correct the use of RecordEvent and simplify recognize_digits.

* MKLDNN conv2 OP kernel added

* TODOs added

* mkldnn conv2d OP refactor

* CanCUDNNBeUsed and CanMKLDNNBeUsed moved

* test cpu float16 data transform

* add isnan etc

* small fix

* fix containsNAN test error

* add data_type transform GPU test

* add float16 GPU example

* fix error

* fix GPU test error

* add context wait

smaller binary proto
avoid untrackable kernel

* Make CUPTI_LIB_PATH not passing by macro.
* Add missing header

* fix nccl version

* enable nccl test

* compile with nccl2

* add ncclGroup; it is necessary in nccl2

* add back libnccl-dev

* move float16 into fluid

* fix include

* move to platform folder