[Done] API for dist train

fix the latex equation in fluid fc layer.

Enforce drop_empty_grad=false When the input of an op is duplicable.

* "remove GPU Sync Interface"

* "fix typo"

* "fix type cast error"

* "fix related Copy with stream"

* "fix failed tests with DevicePool"

* "fix stupid removed position error"

enable alexnet benchmark

use small samples to infer openblas

Fix lr setting of param_attr in Fluid

For input argument with a list of variables, drop_empty_grad is not allowed because it makes the correspondence bewteen a variable and its gradient ambiguous. Use REGISTER_OP_EX to register the op or call InputGrad(?,false) in GradOpDescMaker.

Refine the doc for mul_op and fully connected layer.

Adding documentation for the operators: lod_tensor_to_array , array_to_lod_tensor, create_array, increment (#6807)

* Adding operator assignment

* Adding a prototype for documentation in layers

* small update to re-run Travis

* Removing file from another PR

* Small typo

* Adding documentation for the operators: lod_tensor_to_array , array_to_lod_tensor, create_array, increment

* Fixing indentation issue

* Fixed datatype of input variables

* Polish API docs for assign layer

* Polishing the API docs for concat and assign layer

* Writeup for array write layer

* Fixed the type

fix pip install page 404 links

Rename fill_zeros_like_op's output

* Remove unnecessary reshape in ColwiseSum

Speed up 12s -> 10s.

* Hand write ColwiseAdd in CPU

* add design doc on keys of operator kernel type

* follow comments