enhance test_analyzer to profile ditu inference demo

bug fix

* add is_test for while_op

* Change API

Add scope info in graphviz debug

* dist transpiler add control dependency var between send and recv

* fix async deps

* follow comments and refine

* fix deps connect for rpc ops

refine Node

* Add Python Callstacks when Op::Run error

* Skip op with sub-block

* refactor: refine callstack info's format

* Reshape only support matrix

* Polish Python code

* Fix UT

* Fix Py3

Feature/process lod grad

* Disable in_place in batch_norm API.

* anakin support x86

* fix code style

* add anakin ditu cnn demo

* add timer

* add rnn

* fix inference_anakin_cnn/rnn_test compile error

* make anakin_rnn_tester run

* add anakin_enable_op_time option

* update api/CMakeLists.txt

* enlarge the max_batch_size in anakin.config

* update with comments

fix profiler test

* new doc

* standard

* make fc fuse work with ditu rnn
* add ditu rnn data download to CMAKE

speed up while_op

* doc: append PADDLE_ENFORCE rules to new_op_cn.md

* doc: polish writing

* refactor: polish doc based on advice

 fix ssa bug with batch_norm and  refine the trt

* Fuse Convolution and Eltwise Add into Conv+Bias

* Reduce bias branching at conv_mkldnn_op

* Add MKLDNN build checks for Conv Bias

* Conv-bias: check if bias input exist befor assignment

* Conv-bias: Remove Bias dim check from infershape

It was causing conv3d test to crash upon\ncalling HasInput(Bias)

* link IR graph to analysis graph

* add clean code and update

* add infer_clean_pass

* add ir_pass_manager

* support fc fuse executation

* fix ir circle

Refine  elementwise mul cpu forward

add load slice_vars in io.py

Fix python3 CI job

* add async dist tests

* update delta

* fix transformer test

* refine rmsprop transpile

* update

* fix dist seresnet

only change rpath in Release mode