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.

* Remove unnecessary reshape in ColwiseSum

Speed up 12s -> 10s.

* Hand write ColwiseAdd in CPU

* Rename XXDescBind --> XXDesc

* Fix Compile

* Updating the design doc of Fluid

* Organizing the operator documentation

* Adding a proposed format for operator documentation

* Adding more details to the format

* Move framework.proto to proto namespace

* Fix compile

* Fix compile

* Fix Compile

* "add DeviceContextPool"

* "add devicecontextpool in pybind"

* "add comments in python side "

* "fix static link error"

* "fix CI error"

* "add executor.py"

* "fix CI error"

* "add with gpu macro"

* "remove comment out codes"

* "add TODO items"

* "update init devices"

* add more place_test and rename Cudnn to CUDNN

* fix ci